D&E Technical, Inc., 136 W Main St., Urbana, IL, 61853, USA.
Syngenta Crop Protection, LLC, 410 Swing Rd, Greensboro, NC, 27409, USA.
Environ Monit Assess. 2021 Nov 18;193(12):827. doi: 10.1007/s10661-021-09596-9.
Inclusion of pesticide monitoring data in pesticide risk assessment is important yet challenging for several reasons, including infrequent or irregular data collection, disparate sources procedures and associated monitoring periods, and interpretation of the data itself in a policy context. These challenges alone, left unaddressed, will likely introduce unintentional and unforeseen risk assessment conclusions. While individual water quality monitoring programs report standard operating procedures and quality control practices for their own data, cross-checking data for duplicated data from one database to another does not routinely occur. Consequently, we developed a novel quality control and assurance methodology to identify errors and duplicated records toward creating an aggregated, single pesticide database toward use in ecological risk assessment. This methodology includes (1) standardization and reformatting practices, (2) data error and duplicate record identification protocols, (3) missing or inconsistent limit of detection and quantification reporting, and (4) site metadata scoring and ranking procedures to flag likely duplicate records. We applied this methodology to develop an aggregated (multiple-source), national-scale database for atrazine from a diverse set of surface water monitoring programs. The resultant database resolved and/or removed approximately 31% of the total ~ 385,000 records that were due to duplicated records. Identification of sample replicates was also developed. While the quality control and assurances methodologies developed in this work were applied to atrazine, they generally demonstrate how a properly constructed and aggregated single pesticide database would benefit from the methods described herein before use in subsequent statistical and data analysis or risk assessment.
将农药监测数据纳入农药风险评估很重要，但由于以下几个原因，这具有一定挑战性，包括监测数据的收集频率低或不规律、数据来源、程序和相关监测期不同，以及在政策背景下对数据的解释。这些挑战如果得不到解决，可能会导致风险评估产生非故意和不可预见的结论。虽然个别水质监测计划会报告其自身数据的标准操作程序和质量控制措施，但通常不会对一个数据库中的重复数据与另一个数据库中的数据进行交叉核对。因此，我们开发了一种新颖的质量控制和保证方法，以识别错误和重复记录，从而创建一个汇总的单一农药数据库，用于生态风险评估。该方法包括：(1) 标准化和重新格式化实践；(2) 数据错误和重复记录识别协议；(3) 缺少或不一致的检测限和定量报告；(4) 站点元数据评分和排序程序，以标记可能的重复记录。我们应用此方法来开发一个来自各种地表水监测计划的阿特拉津的汇总（多源）、全国范围的数据库。由此产生的数据库解决和/或删除了大约 31%的总记录，这些记录是由于重复记录造成的。还开发了识别样本重复的方法。虽然本文所开发的质量控制和保证方法适用于阿特拉津，但它们通常可以说明，在用于随后的统计和数据分析或风险评估之前，一个经过适当构建和汇总的单一农药数据库如何从本文所述的方法中受益。
