Interventions to Reduce Pesticide Exposure from the Agricultural Sector in Africa: A Workshop Report.

Despite the fact that several cases of unsafe pesticide use among farmers in different parts of Africa have been documented, there is limited evidence regarding which specific interventions are effective in reducing pesticide exposure and associated risks to human health and ecology. The overall goal of the African Pesticide Intervention Project (APsent) study is to better understand ongoing research and public health activities related to interventions in Africa through the implementation of suitable target-specific situations or use contexts. A systematic review of the scientific literature on pesticide intervention studies with a focus on Africa was conducted. This was followed by a qualitative survey among stakeholders involved in pesticide research or management in the African region to learn about barriers to and promoters of successful interventions. The project was concluded with an international workshop in November 2021, where a broad range of topics relevant to occupational and environmental health risks were discussed such as acute poisoning, street pesticides, switching to alternatives, or disposal of empty pesticide containers. Key areas of improvement identified were training on pesticide usage techniques, research on the effectiveness of interventions targeted at exposure reduction and/or behavioral changes, awareness raising, implementation of adequate policies, and enforcement of regulations and processes.

Erosion rills offset the efficacy of vegetated buffer strips to mitigate pesticide exposure in surface waters.

Regulatory risk assessment considers vegetated buffer strips as effective risk mitigation measures for the reduction of runoff-related pesticide exposure of surface waters. However, apart from buffer strip widths, further characteristics such as vegetation density or the presence of erosion rills are generally neglected in the determination of buffer strip mitigation efficacies. This study conducted a field survey of fruit orchards (average slope 3.1-12.2%) of the Lourens River catchment, South Africa, which specifically focused on the characteristics and attributes of buffer strips separating orchard areas from tributary streams. In addition, in-stream and erosion rill water samples were collected during three runoff events and GIS-based modeling was employed to predict losses of pesticides associated with runoff. The results show that erosion rills are common in buffer strips (on average 13 to 24 m wide) of the tributaries (up to 6.5 erosion rills per km flow length) and that erosion rills represent concentrated entry pathways of pesticide runoff into the tributaries during rainfall events. Exposure modeling shows that measured pesticide surface water concentrations correlated significantly (R(2)=0.626; p<0.001) with runoff losses predicted by the modeling approach in which buffer strip width was set to zero at sites with erosion rills; in contrast, no relationship between predicted runoff losses and in-stream pesticide concentrations were detected in the modeling approach that neglected erosion rills and thus assumed efficient buffer strips. Overall, the results of our study show that erosion rills may substantially reduce buffer strip pesticide retention efficacies during runoff events and suggest that the capability of buffer strips as a risk mitigation tool for runoff is largely overestimated in current regulatory risk assessment procedures conducted for pesticide authorization.

Combined effects of discharge, turbidity, and pesticides on mayfly behavior: experimental evaluation of spray-drift and runoff scenarios.

The effects of the pyrethroid-insecticide cypermethrin (CYP), increased flow speed (Flow), and increased suspended particles (Part) on drift behavior and activity of mayfly nymphs (Baetis harrisoni) were investigated both individually and in combination in a laboratory stream microcosm. Spray-drift trials were performed by exposing the nymphs to 1 microg/L of CYP. During runoff trials (CYP x Part), contaminated sediment containing 2,000 microg/kg of CYP was introduced to the microcosm at a concentration of 500 mg/L. Both trials were carried out under high-flow (CYP x Flow and CYP x Part x Flow) and low-flow (CYP and CYP x Part) conditions, and for all cases, control experiments were performed. Drift rate, drift density (for any treatments with increased flow), and activity were used as behavioral endpoints. Multifactorial analysis of variance shows that CYP exposure significantly increased the drift, whereas Part and Flow trials significantly decreased the drift (p < 0.05). In addition, activity decreased significantly under high-flow conditions. The CYP x Part and CYP x Flow treatments resulted in increased drift rate and drift density, respectively, whereas Part x Flow and CYP x Part x Flow treatments resulted in decreased drift density. The CYP x Part and CYP x Flow trials had a significant antagonistic, interactive effect on drift rate and drift density, respectively, with measured levels being lower than expected levels. The reduction in bioavailability of CYP in the presence of increased flow and sediment levels suggests that mayflies are more likely to be affected by spray-drift exposure (CYP) than by runoff exposure (CYP x Part x Flow). Results indicate that mayflies reacted actively in response to flow conditions and passively in response to pesticide exposure.

Validation of the AGDISP model for predicting airborne atrazine spray drift: A South African ground application case study.

Air dispersion software models for evaluating pesticide spray drift during application have been developed that can potentially serve as a cheaper convenient alternative to field monitoring campaigns. Such models require validation against field monitoring data in order for them to be employed with confidence, especially when they are used to implement regulatory measures or to evaluate potential human exposure levels. In this case study, off-target pesticide drift was monitored during ground application of a pesticide mixture to a sorghum field in South Africa. Atrazine was used as a drift tracer. High volume air sampling onto polyurethane foam (PUF) was conducted at six downwind locations and at four heights at each sampling point. Additional data, including meteorological information, required to simulate the spray drift with the AGDISP® air dispersion model was collected. The PUF plugs were extracted by a plunger method utilizing a hexane:acetone mixture with analysis by GC-NPD (94.5% recovery, 3.3% RSD, and LOD 8.7 pg). Atrazine concentrations ranged from 4.55 ng L(-1) adjacent to the field to 186 pg L(-1) at 400 m downwind. These results compared favourably with modeled output data, resulting in the validation of the model up to 400 m from the application site for the first time. Sensitivity studies showed the importance of droplet size distribution on spray drift, which highlighted the need for good nozzle maintenance. Results of this case study indicate that the model may provide meaningful input into environmental and human health risk assessment studies in South Africa and other developing countries.

Predicting runoff-induced pesticide input in agricultural sub-catchment surface waters: linking catchment variables and contamination.

An urgent need exists for applicable methods to predict areas at risk of pesticide contamination within agricultural catchments. As such, an attempt was made to predict and validate contamination in nine separate sub-catchments of the Lourens River, South Africa, through use of a geographic information system (GIS)-based runoff model, which incorporates geographical catchment variables and physicochemical characteristics of applied pesticides. We compared the results of the prediction with measured contamination in water and suspended sediment samples collected during runoff conditions in tributaries discharging these sub-catchments. The most common insecticides applied and detected in the catchment over a 3-year sampling period were azinphos-methyl (AZP), chlorpyrifos (CPF) and endosulfan (END). AZP was predominantly found in water samples, while CPF and END were detected at higher levels in the suspended particle samples. We found positive (p < 0.002) correlations between the predicted average loss and the concentrations of the three insecticides both in water and suspended sediments (r between 0.87 and 0.94). Two sites in the sub-catchment were identified as posing the greatest risk to the Lourens River mainstream. It is assumed that lack of buffer strips, presence of erosion rills and high slopes are the main variables responsible for the high contamination at these sites. We conclude that this approach to predict runoff-related surface water contamination may serve as a powerful tool for risk assessment and management in South African orchard areas.

Predicted and measured levels of azinphosmethyl in the Lourens River, South Africa: comparison of runoff and spray drift.

Runoff and spray drift are important sources of nonpoint pesticide pollution in surface waters, but few studies have directly compared these routes of input in an exposure assessment scenario. To this end, a runoff formula suggested by the Organization for Economic and Cooperative Development (Paris, France) and basic drift values (95th percentiles) were integrated into a geographical information system (GIS) to predict runoff and spray drift-related loading of azinphosmethyl (AZP) in the Lourens River (LR), South Africa. The GIS-integrated calculations were first validated in the tributaries of the river, where measured loads were well predicted for both runoff (r(20 = 0.95; p < 0.0001; n = 9) and spray drift (r(2) = 0.96; p = 0.0006; n = 8). Through extrapolation to the catchment scale containing 400 ha of orchards, the GIS-integrated calculations predicted similar loads of AZP as measured in the Lourens River mainstream for six runoff (between a factor of 1.03 and 1.86 lower) and six spray drift (between a factor of 1.1 and 2.4 higher) events. Mean measured loads per event were significantly (p = 0.004) higher for runoff (27.8 +/- 19.1 g) than for spray drift (0.69 +/- 0.32 g). Based on long-term meteorological data and average application regimes, runoff leads to a higher annual load (47.6 g) than spray drift (5.5 g) in the Lourens River. Runoff is clearly a more important source of nonpoint pollution in the studied catchment, and mitigation strategies should focus first on addressing this aspect on a catchment scale and second on addressing problem areas on a subcatchment scale.

A combined microcosm and field approach to evaluate the aquatic toxicity of azinphosmethyl to stream communities.

We evaluated the potential effects of the organophosphate insecticide azinphosmethyl (AZP) in a combined microcosm and field approach. The upper regions of the Lourens River, South Africa, are free of contamination (control site), whereas the subsequent stretches flowing through a 400-ha orchard area receive transient insecticide pollution (e.g.. 0.82 microg/L AZP, 344 microg/kg chlorpyrifos) following spray drift and runoff (contaminated site). Stones taken from the control site were transferred to outdoor microcosms (1.5 x 0.2 x 0.2 m), providing 12 core species and approximately 350 individuals per microcosm. Microcosms were contaminated for 1 h with AZP (control, 0.2, 1, 5, and 20 microg/L; three replicates each), and acute effects on survival were evaluated 6 d following exposure. The two strongest treatments (measured concentrations: 19.2 +/- 1.0 and 4.9 +/- 0.3 microg/L, respectively) resulted in a significantly (analysis of variance) reduced invertebrate density, attributed mainly to various insect taxa, such as Demoreptus sp., Castanophlebia sp., Simuliidae, and Chironomidae. In contrast, Aeshna sp., Dugesia sp., Ceratopogonidae, and Cheumatopsyche sp. were unaffected. In parallel, we conducted a quantitative macroinvertebrate survey at the control site and the contaminated site of the Lourens River after the seasonal pesticide application period. The two sites contained a similar number of species but differed considerably in their species composition and abundances. Five of the eight species that were affected by AZP in the microcosm study occurred in the field at significantly lower densities at the contaminated than at the control site or were absent at the contaminated site. All of the four species that were unaffected in the microcosm occurred at significantly higher densities at the contaminated field site. Only 3 of the 12 species reacted differently in the microcosm and the field study. We conclude that microcosm studies employing a field-relevant design could be linked successfully to field studies and our results suggest that transient pesticide contamination affects the aquatic communities of the Lourens River.

Prioritizing agricultural pesticides used in South Africa based on their environmental mobility and potential human health effects.

South Africa is the largest user of pesticides in sub-Saharan Africa and many studies have highlighted the occurrence of pesticides in water resources. Poor management of water treatment facilities in combination with a relatively high dependency on untreated water from boreholes and rivers creates the potential for exposure of human communities to pesticides and their associated health effects. Pesticide use, physicochemical and toxicity data was therefore used to prioritize pesticides in terms of their potential risk to human health. After eliminating pesticides used in very low quantities, four indices were used to prioritize active ingredients applied in excess of 1000 kg per annum; the quantity index (QI) which ranked pesticides in terms of the quantity of their use; the toxicity potential index (TP) which ranked pesticides according to scores derived for their potential to cause five health effects (endocrine disruption, carcinogenicity, teratogenicity, mutagenicity and neurotoxicity); hazard potential index (HP) which multiplied the TP by an exposure potential score determined by the GUS index for each pesticide (to provide an indication of environmental hazard); and weighted hazard potential (WHP), which multiplied the HP for a pesticide by the ratio of its use to the total use of all pesticides in the country. The top 25 pesticides occurring in each of these indices were identified as priority pesticides, resulting in a combined total of 69 priority pesticides. A principal component analysis identified the indices that were most important in determining why a specific pesticide was included in the final priority list. As crop specific application pesticide use data was available it was possible to identify crops to which priority pesticides were applied to. Furthermore it was possible to prioritize crops in terms of the specific pesticide applied to the crop (by expressing the WHP as a ratio of the total amount of pesticide applied to the crop to the total use of all pesticides applied in the country). This allows for an improved spatial assessment of the use of priority pesticides. The methodology applied here provides a first level of basic, important information that can be used to develop monitoring programmes, identify priority areas for management interventions and to investigate optimal mitigation strategies.

Development and field validation of an indicator to assess the relative mobility and risk of pesticides in the Lourens River catchment, South Africa.

A GIS based pesticide risk indicator that integrates exposure variables (i.e. pesticide application, geographic, physicochemical and crop data) and toxicity endpoints (using species sensitivity distributions) was developed to estimate the Predicted Relative Exposure (PREX) and Predicted Relative Risk (PRRI) of applied pesticides to aquatic ecosystem health in the Lourens River catchment, Western Cape, South Africa. Samples were collected weekly at five sites from the beginning of the spraying season (October) till the beginning of the rainy season (April) and were semi quantitatively analysed for relevant pesticides applied according to the local farmers spraying programme. Monitoring data indicate that physicochemical data obtained from international databases are reliable indicators of pesticide behaviour in the Western Cape of South Africa. Sensitivity analysis identified KOC as the most important parameter influencing predictions of pesticide loading derived from runoff. A comparison to monitoring data showed that the PREX successfully identified hotspot sites, gave a reasonable estimation of the relative contamination potential of different pesticides at a site and identified important routes of exposure (i.e. runoff or spray drift) of different pesticides at different sites. All pesticides detected during a monitored runoff event, were indicated as being more associated with runoff than spray drift by the PREX. The PRRI identified azinphos-methyl and chlorpyrifos as high risk pesticides towards the aquatic ecosystem. These results contribute to providing increased confidence in the use of risk indicator applications and, in particular, could lead to improved utilisation of limited resources for monitoring and management in resource constrained countries.

Fate and effects of azinphos-methyl in a flow through wetland in South Africa.

Our knowledge about the effectiveness of constructed wetlands in retaining agricultural nonpoint-source pesticide pollution is limited. A 0.44-ha vegetated wetland built along a tributary of the Lourens River, Western Cape, South Africa, was studied to ascertain the retention, fate, and effects of spray drift-borne azinphos-methyl (AZP). Composite water samples taken at the inlet and outlet during five spray drift trials in summer 2000 and 2001 revealed an overall reduction of AZP levels by 90 +/- 1% and a retention of AZP mass by 61 +/- 5%. Samples were collected at the inlet outlet, and four platforms within the wetland to determine the fate and effect of AZP in the wetland after direct spray drift deposition in the tributary 200 m upstream of the inlet. Peak concentrations of AZP decreased, and the duration of exposure increased from inlet (0.73 microg/L; 9 h) via platforms 1 and 4 to outlet (0.08 microg/L; 16 h). AZP sorbed to plants or plant surfaces, leading to a peak concentration of 6.8 microg/kg dw. The living plant biomass accounted for 10.5% of the AZP mass initially retained in the wetland, indicating processes such as volatilization, photolysis, hydrolysis, or metabolic degradation as being very important AZP was not detected in sediments. Water samples taken along two 10-m transects situated perpendicular to the shore indicated a homogeneous horizontal distribution of the pesticide: 0.23 +/- 0.02 and 0.14 +/- 0.04 microg/L (n = 5), respectively. Both Copepoda (p = 0.019) and Cladocere (p = 0.027) decreased significantly 6 h postdeposition and remained at reduced densities for at least 7 d. In parallel, the chlorophyll a concentration showed an increase, although not significant, within 6 h of spray deposition. The study highlights the potential of constructed wetlands as a risk-mitigation strategy for spray drift-related pesticide pollution.