Use of different passive sampling approaches for a comprehensive and time-integrated sampling of pesticides in tropical streams in a vegetable growing area

There is a newer version of this dataset.

This package contains the supplementary information (SI) of chapter 2 of the dissertation of Frederik T. Weiss with the Dissertation No. ETH 27434 (defended: 24th February, 2021), entitled: "Pesticides in a tropical Costa Rican stream catchment: from monitoring and risk assessment to the identification of possible mitigation options". Generally within this thesis the supplementary information (SI) is divided into three parts (SI A, SI B, SI C). For each chapter, SI A section contains background information/data for the reader with quick and easy access added directly after each main chapter. SI B contains raw data, further processed data for analysis, and figures of processed data presented as Excel files. SI C combines the R scripts with information and commands utilized for the statistical analysis.

The abstract of chapter 2 reads as follows:

"For monitoring of pesticides in tropical streams, cost-efficient and easily applicable approaches are needed. Moreover, to capture short pesticide concentration peaks, a time-integrated sampling is preferable to conventional snapshot grab sampling. Passive sampling approaches fulfil these criteria. Therefore, this chapter focusses on the application of three passive sampling devices to monitor 275 pesticides and pesticide transformation products (PPTP) in the horticultural Tapezco river catchment over several months in two consecutive years. Two of the samplers were sorbent-based: reverse phase sulfonated styrene-divinylbenzene (SDB) disks and polydimethylsiloxane (PDMS) sheets, yielding biweekly integrated averaged PPTP concentrations. The third sampler was a low-cost, non-sorbent-based, water level proportional sampling system (WLPSS), yielding water level-weighted, biweekly integrated PPTP concentrations. The objectives were to (1) test the performance and robustness of these samplers (2) obtain comprehensive quantitative pesticide concentration data and (3) provide recommendations for their field application in future monitoring campaigns. Of the 275 targeted PPTP, 87 polar and semi-polar PPTP were detected with the SDB method and 99 with the WLPSS, of which 77 were found with both systems. In several cases (10 with SDB, 22 with WLPSS), a pesticide was only detected by one of the set-ups; this exclusive detection could be due to the respective substance concentrations being close to or below the method limit of quantification (MLOQ) for the sampler where it was not detected. Despite the different sampling principles for SDB and WLPSS, the same pesticides (carbendazim and flutolanil) were found with the highest median water concentrations (> 100 ng/L) with both samplers. The complementary PDMS system allowed detection of 11 non-polar pesticides. Among these, cypermethrin, chlorpyrifos and permethrin showed the highest concentrations (> 2 ng/L). Chlorpyrifos was the only pesticide detected with all three sampling techniques. Standard deviations for detected chlorpyrifos concentrations were the highest for SDB sampling, likely due to a lag-phase in sampling across the membrane covering the sampler due to the chemical’s high hydrophobicity. Moreover, derived chlorpyrifos water concentrations were significantly higher using the WLPSS compared to SDB and PDMS sampling. This was also seen for another six pesticides sampled with the WLPSS compared to SDB sampling. Higher concentrations detected via WLPSS can be explained by the ability of the WLPSS to collect pesticide peaks associated with heavy rainfall events and linked to rise of water levels in a more pronounced fashion as compared to the time-integrated sampling manner of the SDB and PDMS samplers. Yet, only a small portion, 15%, of the WLPSS samples collected, could be used to yield water level-weighted, time-integrated concentration (CWLW) data, calling for a need to further optimize and standardize the application of this device. Of the devices tested, the SDB disks were the easiest to apply and the most cost-efficient for short-term monitoring campaigns. The SDB sampling can be conducted in sparsely equipped laboratory facilities, while for the PDMS sheets and the WLPSS, sample preparation and extraction are technically more demanding."