Identification of pesticide input pathways in tropical streams as a basis to propose potential mitigation options

This package contains the supplementary information (SI) of chapter 4 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 4 reads as follows:

"Finding targeted strategies to mitigate entry of pesticides into surface waters in areas of intense agriculture is challenging. This holds especially true in little studied areas with very distinct topographic characteristics and unconventional field cultivation practices, such as in the tropical Tapezco river catchment in Costa Rica. Within this catchment, areas with steep slopes are used for intense horticultural farming of mainly vegetables. This is exclusively done by a farming practice similar to contour farming, the practice of tilling land with furrows along parallel lines of consistent elevation in order to conserve rainwater and to prevent soil losses by erosion. At the same time, slope-directed paths are implemented to act as drainage system to avoid stagnant water on the fields during heavy rain events, though as well connecting the fields directly with the streams, which enable a fast pesticide transport. Indeed, a significant contamination of streams with pesticides and pesticide transformation products (PPTP) throughout the Tapezco river catchment has been confirmed, leading to considerable toxicological risks to aquatic communities, urgently calling for effective mitigation strategies to reduce PPTP inputs. To identify how PPTP are transported from horticultural areas into streams of the Tapezco river catchment, different PPTP transportation pathways were considered. The first investigated pathway was via handling practices of pesticides by farmers and field workers, where inappropriate handling was proposed to lead to sporadically distributed pesticide inputs unrelated to hydrology. The second studied pathway was surface run-off. Typically, heavy precipitation events are found to be important drivers for the surface-based transport of pesticides into the streams. Thus, such pesticide inputs can be assumed to correlate positively with water levels in the receiving streams. Surface run-off is additionally favored by the slope-directed paths on the fields, which directly connect fields with the streams. Therefore, the influence of prevalent topographical and hydrological variables on PPTP inputs via surface run-offs were studies within this thesis. The third potential investigated input pathway was the leaching of pesticides into the ground from where pesticides can enter streams via exfiltration through river banks. This path would be expected to lead to a constant input that is negatively correlated with water levels. To investigate the role of these pathways in transporting PPTP into the streams, pesticide peaks unrelated to hydrology were identified based on measured environmental concentrations (MEC) of PPTP and compared with water level time series. Survey data about pesticide handling practices were evaluated additionally. Temporal PPTP distributions were investigated during three sampling periods (ΔT1, Δ2a, Δ2b) within 2015 and 2016 and spatial trends were studied at eight sub-catchment (SC) sites. In addition, knowledge on the topography (share of horticultural land, share of forest in the 100 m stream buffer zone, average slopes of the horticultural fields) and hydrology (median water level factors) was considered. These variables were referred to as explanatory variables while 20-, 50- and 80-percentiles of MEC were considered dependent variables. The explanatory and dependent variables were correlated via linear regression modelling for identifying the most important determinants of PPTP transport. There, 20-percentiles represent a scenario with low precipitations, no or low surface run-offs and low PPTP inputs; 50-percentiles a scenario with medium precipitations, resulting in medium surface run-offs and PPTP inputs; and 80-percentiles a scenario with high precipitations, heavy surface run-offs and high PPTP inputs into streams. With a focus on potential mitigation measures achieving the highest effectiveness for reducing risks to aquatic biota, analyses were performed on a sub-set of PPTP that dominated the risks to aquatic organisms, along with three transformation products (TP) to calculate TP/PPTP ratios as a measure of pesticide residence time. The correlation analysis of the PPTP input pathways was again based on eight SC sites. The input of three pesticides were very likely due to inappropriate handling. For five additional pesticides, the input via inappropriate handling seemed probable. Temporal exposure trends were observed by comparing the MEC during the sampling period with reduced precipitation (ΔT1, in 2015) with the MEC detected at periods with normal precipitations (Δ2a, Δ2b, in 2016). In addition, spatial trends were investigated by conducting a cluster analysis with the MEC PPTP data (20-, 50- and 80-percentiles) among the different sites. Particularly the pesticide distributions at SC2 and SC3 were different compared to other sites (SC1, SC4, SC6, SC7 and SC8). However, except for the 20-percentile scenario, the pesticide distribution at SC5 was similar compared to that at SC2 and SC3, forming one sub-cluster. Linear regression models helped to find relationships between two explanatory variables, namely, the share of forest in the buffer zone, and mean slopes of horticultural fields, and the dependent variable, MEC percentiles in streams. For five PPTP, boscalid, diazinon, diuron-desdimethyl, linuron and prometryn + terbutryn the percentile concentrations decreased significantly with increasing share of forest in 100 m river buffer zone considering all scenarios. With regard to the horticultural mean slope, for cyhalothrin and thiamethoxam, the percentile concentrations increased with increasing mean slopes of the horticultural areas for all three scenarios. A high share of forest in the buffer zone worked generally as barrier for input via surface run-off, but not for all PPTP. For the fungicide, carbendazim, increased average slopes did not favor the input into the streams and inputs were low even at sites with horticultural areas with a high mean slope (80 percentile scenario). By analyzing groundwater samples it became apparent that, especially in SC with horticultural fields with low average slopes, a leaching of PPTP into groundwater and further transport into the streams via exfiltration might be possible. Based on this assessment, three avenues for mitigating input of PPTP into the streams could be deduced: to provide training workshops for better handling as well as biobeds for proper disposal; to avoid cultivation of crops in high need insecticides on steep slopes; and to establish forested buffer zones between the fields and the streams."

Dataset extent

Data and Resources


Weiss, F., Schaub, N., Wey, H., Eggen, R., & Stamm, C. (2021). Identification of pesticide input pathways in tropical streams as a basis to propose potential mitigation options (Version 1.0) [Data set]. Eawag: Swiss Federal Institute of Aquatic Science and Technology.


Open Data Open Data
  • Weiss, Frederik
  • Schaub, Niccolà
  • Wey, Hannah
  • Eggen, Rik
  • Stamm, Christian
Keywords short cuts,surface runoff,land use,topography,LC-MS,GC-MS,pyrethroids,pathways,groundwater,pesticide mitigation
  • concentration
  • precipitation
  • waterlevel
Substances (scientific names)
  • 2,6-Dichlorbenzamide C7H5Cl2NO
  • Acephate C4H10NO3PS
  • Acetochlor-ESA + Alachlor-ESA C14H21NO5S
  • Bifenthrine C23H22ClF3O2
  • Boscalid C18H12Cl2N2O
  • Bromoxynil C7H3Br2N1O1
  • Carbendazim C9H9N3O2
  • Carbofuran C12H15NO3
  • Chloridazon-desphenyl C4H4ClN3O
  • Chlorothalonil-4-hydroxy-carbonacid amide C8H3Cl3N2O2
  • Chlorpyrifos C9H11Cl3NO3PS
  • Compound Molecular formula
  • Cyhalothrin C23H19ClF3NO3
  • Cypermethrin C22H19Cl2NO3
  • Cyromazin C6H10N6
  • Deltamethrin C22H19Br2NO3
  • Diazinon C12H21N2O3P1S1
  • Dimethoate C5H12NO3PS2
  • Dimethomorph C21H22ClNO4
  • Diuron C9H10Cl2N2O1
  • Diuron-desdimethyl C7H6Cl2N2O
  • Diuron-desmonomethyl C8H8Cl2N2O
  • Fipronil C12H4Cl2F6N4O1S1
  • Imidacloprid C9H10ClN5O2
  • Iodopropynyl butyl-carbamate (IPBC) C8H12I1N1O2
  • Linuron C9H10Cl2N2O2
  • Metribuzin C8H14N4O1S1
  • Metribuzin-Desamino (DA) C8H13N3OS
  • Permethrin C21H20Cl2O3
  • Prometryn + Terbutryn C10H19N5S
  • Propamocarb C9H20N2O2
  • Tebuconazole C16H22ClN3O
  • Thiamethoxam C8H10ClN5O3S
Substances (generic terms)
  • fungicide
  • herbicide
  • insecticide
  • metabolite
  • transformation product
  • groundwater
  • river
  • stream
  • 2015-07 TO 2015-10
  • 2016-05 TO 2016-10
Geographic Name(s)
  • Tapezco, Costa Rica
Review Level general
Curator Weiss, Frederik
DOI 10.25678/0004R4