Atrazine in the environment 20 years after its ban : long-term monitoring of a shallow aquifer (in western Germany) and soil residue analysis

Vonberg, David; Rüde, Thomas R. (Thesis advisor); Vanderborght, Jan (Thesis advisor)

Jülich : Forschungszentrum Jülich GmbH, Zentralbibliothek (2015, 2016)
Book, Dissertation / PhD Thesis

In: Schriften des Forschungszentrums Jülich. Reihe Energie & Umwelt = Energy & environment 293
Page(s)/Article-Nr.: 1 Online-Ressource (149 Seiten) : Illustrationen, Diagramme, Karten

Dissertation, RWTH Aachen, 2015


Atrazine, one of the worldwide most widespread herbicides, was banned in Germany in 1991 and in the European Union in 2004, due to findings of atrazine concentrations in ground- and drinking waters exceeding the threshold value of 0.1 µg L-1. Nevertheless atrazine and the metabolite deethylatrazine were still detected in German aquifers more than 10 years after its prohibition, often without any considerable decreasing trend in groundwater concentration. Because atrazine was already found to be persistent in soils for more than two decades after the last application, the hypothesis was raised that a continued release of atrazine residues from the soil into groundwaters might sustain atrazine groundwater concentrations on elevated levels. The overall objective of this study was to investigate the occurrence and concentration trends of atrazine and its main metabolites in the groundwater-soil environment after the prohibition of its use. Accordingly, in this study results of i) 20 years of atrazine groundwater monitoring of a a shallow aquifer in western Germany since its ban and ii) atrazine soil residue analyses in the vadose zone of the same study area 21 years after its ban are presented. The phreatic Zwischenscholle aquifer located in the Lower Rhine Embayment is exposed to intensive agricultural land use and is highly susceptible to contaminants due to a shallow water table. In total 60 observation wells (OWs) have been monitored since 1991, of which 11 are sampled monthly today. Descriptive statistics of monitoring data were derived using the "regression on order statistics" (ROS) data censoring approach, estimating values for nonquantifiable values rather than substitute them by e.g. half of the limit of quantification and taking the risk of biasing statistical parameters. The monitoring data shows that even 20 years after the ban of atrazine, the groundwater concentrations of sampled OWs remain on a level close to the threshold value of 0.1 µg L-1 without any considerable decrease. The spatial distribution of atrazine concentrations is highly heterogeneous with OWs exhibiting permanently concentrations above the regulatory threshold on the one hand and other OWs with concentrations mostly below the limit of quantification (LOQ) on the other hand. Here atrazine concentrations show upward, downward or approximately constant trends. The deethylatrazine-to-atrazine ratio (DAR) was used to distinguish between diffuse - and point-source contaminations. A DAR around unity (slightly smaller for thin vadose zones like for the investigated aquifer) suggests a contamination of an aquifer by diffuse pathways, resulting in significant metabolization of atrazine to deethylatrazine due to a longer contact time to soil microorganisms. Conversely, point-source contaminations where the contaminant enters the aquifer directly by e.g. macropore flow results in negligible deethylation and hence a DAR close to zero. A global mean DAR value of 0.84 for the monitoring data of the Zwischenscholle aquifer indicates mainly diffuse contamination. Also most of the DARs for single observation wells suggest mainly diffuse pollution, except for one OW with a mean DAR of 0.02, clearly indicating point-source contamination. Principle Component Analysis (PCA) of the monitoring dataset demonstrated relationships between the metabolite deisopropylatrazine and its parent compound simazine but not with atrazine, and deethylatrazine, atrazine, nitrate, and the specific electrical conductivity. These parameters indicate diffuse agricultural impacts on groundwater quality. The groundwater monitoring findings point at the difficulty to estimate mean concentrations of contamination for entire aquifers and to evaluate groundwater quality based on average parameters. However, analytical data of monthly sampled single observation wells provide adequate information to characterize local contamination and evolutionary trends of pollutant concentration. For atrazine soil residue analysis three soil cores reaching down to the groundwater table (approximately 3 m below soil surface) were taken in an agricultural field where atrazine was applied prior to its ban. It is uncertain if atrazine was applied in total two or three times with a recommended dose of 0.96 kg ha-1. Eight layers were separated (0-10 cm, 10-30 cm, 30-60 cm, 60-100 cm, 100-150 cm, 150-200 cm, 200-250 cm, 250-300 cm) for atrazine residue analysis and soil parameters (grain size distribution, pH, cation exchange capacity (CECeff) and organic carbon content). Soil samples of each layer were extracted using accelerated solvent extraction (ASE) and analyzed by LC-MS/MS analysis. Prior to this analysis, a method validation was conducted to find optimum extraction parameter combinations. For all extractions a methanol/water (4:1, v/v) solvent was used. The highest quantifiable atrazine extraction yield amongst all extraction parameter combinations between 100°C and 135°C and 100 bar, 150 bar and 207 bar was obtained for 100°C and 207 bar. Atrazine yields were generally higher for higher pressures. Possibly the higher pressure facilitates soil matrix penetration by the solvent. Extractions using 135°C and the highest pressure of 207 bar resulted in quantified concentration of atrazine 31 % lower than those using 100°C. Probably, the higher extraction temperature lead to an increased co-extraction of soil-matrix compounds, which caused a quenching effect and hence less quantifiable atrazine. Extracted atrazine concentrations of the different layers of the soil cores ranged between 0.2 µg kg-1 and 0.01 µg kg-1 for topsoil and subsoil, respectively. Averaged residual atrazine accounts for 0.01 % of the applied mass in the top layer and 0.07 % in the total soil profile (for in total 3 applications). However, the calculation can only be treated as a conservative estimate, because spatial information of atrazine field applications and the correct number of applications (2 or 3 times) are not available. A complete and instantaneous remobilization of atrazine residues from the unsaturated zone, leaching to and mixing with the entire groundwater body would result in a mean groundwater concentration of 0.002 µg L-1. In contrast, considering local atrazine groundwater contamination below an atrazine residue area by a complete instantaneous remobilization of the latter and vertical mixing with the groundwater body below, atrazine groundwater concentrations would be 0.068 µg L-1. Based on the first scenario, long term leaching of aged atrazine residues from the vadose zone seems to marginally contribute to sustaining average groundwater concentrations of the Zwischenscholle aquifer, which remained constantly close to the threshold limit of 0.1 µg L-1 even 20 years after the ban of atrazine. In contrast, the second scenario shows that ongoing local leaching of atrazine from soil residues might result in locally elevated atrazine groundwater concentrations, what might be reflected by the high spatial variability in atrazine groundwater concentrations in the investigated aquifer. A conservative estimate suggests an atrazine half-life value of approximately 2 years for the soil zone, which is significantly higher than the highest atrazine half-life values found in literature (433 days [1.19 years] for top soil). This value only can be taken as rough orientation and most probably underestimates the atrazine half-life time in this soil, because i) non-extractable atrazine could not be included in the calculation and ii) the first two applications were executed before 1991 (information of the exact time of application is missing) and iii) for aged atrazine residues and increased resistance to biodegradation, atrazine degradation in soils rather follows multi-rate kinetics than assumed first order kinetics, what could result in an overestimation of decay rates. These findings show that atrazine persistence in the field might be distinctively higher than predicted assuming first-order degradation kinetics and using half-life values obtained from lab experiments which reach a maximum of 433 days for topsoils. Generally, literature values for the organic carbon normalized distribution coefficient (KOC) and dissipation half-life value for atrazine show a wide range between 25 and 600 L Kg 1 and a few days up to 433 days. Until now there is a lack of the understanding of how herbicide degradation rates vary according to spatial heterogeneity of soil properties. Furthermore, important determining factors influencing degradation like microbial ecology and its spatial variability have been neglected so far for pesticide fate predictions. Accordingly, the accuracy of model predictions of catchment scale atrazine behavior on the long-term based on first-order-kinetics and standard laboratory-derived sorption parameter values may be not reliable. Thus the risk of long-term adverse environmental effects may be higher than estimated. In consequence, there is a need for more realistic pesticide risk assessments and regulation procedures besides standard models for pesticide fate predictions. Finally, considering the key finding that the persistence of particular pesticides in groundwater may be highly underestimated by pesticide fate predictions based on laboratory short-term studies, contaminant monitoring in the groundwater-soil environment remains of highest importance, to i) detect potential groundwater contaminations, ii) re-consider pesticide fate predictions, iii) limit or ban the use of contaminants frequently exceeding thresholds and iv) treat drinking water adequately.