This readme file provides metadata for tables 1 through 9 for U.S. Geological Survey Scientific Investigations Report 2020–5116, "Quality of data from the U.S. Geological Survey National Water Quality Network for Water Years 2013-17" by Laura Medalie and Laura M. Bexfield.
Metadata include table titles, headnotes, abbreviations, and explanations of column names and column contents.

Table 1. Information about chemical compounds in water samples from the U.S. Geological Survey (USGS) National Water Quality Network analyzed at the USGS National Water Quality Laboratory for water years 2013–17.
Headnote: All compounds except for suspended sediment are analyzed at the U.S. Geological Survey National Water Quality Laboratory; —, not applicable.
Explanation of column names and contents
Compound group: N, nutrients; C, carbon; P, physical; S, sediment; M, major ion; T, trace element; Pe, pesticide or degradate
CASRN: Chemical Abstracts Service Registry Number, a registered trademark of the American Chemical Society. CAS recommends the verification of the CASRNs through CAS Client Services. 
Unit: mg/L, milligram per liter; u/cm, unit per centimeter; µg/L, microgram per liter; ng/L, nanogram per liter; NTRU, nephelometric turbidity ratio unit; µS/cm, microsiemens per centimeter at 25 degrees Celsius; NTU, nephelometric turbidity unit
Date range, DL, RL, RL code (1): Date range, concentration of detection limit, concentration of reporting level, or reporting level code applicable to start of study period. Date format is yyyymmdd.
Date range, DL, RL, RL code (2): Date range, concentration of detection limit, concentration of reporting level, or reporting level code applicable to end of study period. Date format is yyyymmdd.  
   For potassium, from 20170525 to 20170930, the DL was 0.1 mg/L, the RL was 0.2 mg/L, and the RL code was detection limit calculated by DQCALC software (dldqc).
   For arsenic, from 20161001 to 20170930, the DL was 0.05 mg/L, the RL was 0.1 mg/L, and the RL code was detection limit by blank data (dlblk). 
   For iron, from 20170525 to 20170930, the DL was 10 mg/L, the RL was20 mg/L, and the RL code was dldqc.
   For strontium, from 20161001 to 20170930, the DL was 0.5 mg/L, the RL was 1 mg/L, and the RLcode was dlblk.
RL code (1) and (2):  ltmdl, long-term method detection limit; mdl, method detection limit; mrl, minimum reporting level; irl, interim reporting level; rldqc, reporting limit established based on detection limit calculated by DQCALC software; /, or.
VQC: Validation quality code, —, no qualification, E, permanent estimated remark code (Sandstrom and others, 2015), E (2016), estimated remark code that went into effect January 1, 2016


Table 2. Quality-control datasets and associated data-quality objectives for water-quality sampling in the U.S. Geological Survey National Water Quality Network and repositories for interpreted quality-control results for water years 2013–17.
Headnote: For the overlapping period between water years (WYs) 2013 and 2015, pesticide data for field replicates in Medalie and Shoda (2020) repeat pesticide data for field replicates from Shoda and others (2017a). FMS, field matrix spike; LB, laboratory blank; LS, laboratory spike; NWQL, National Water Quality Laboratory; QSB, Quality Systems Branch; IBW, inorganic blank water; OBW, organic blank water; QC, quality control.
Explanation of column names and contents
Source of QC sample: NWQL, National Water Quality Laboratory; QSB, Quality Systems Branch
Used to assess: DF, detection frequency.
Data-quality objective: %, percent; UCL, upper confidence limit; DF, detection frequency; RSD, relative standard deviation; ?, greater than or equal to; ?, less than or equal to; >, greater than.
Violations of data-quality objective: Identified in report, years and compounds with this potential bias are identified in this report.
Data repository applicable compound group: N, nutrients; C, carbon; P, physical; S, sediment; M, major ion; T, trace element; Pe, pesticide or degradate


Table 3. Summary information for field blanks and surface-water samples of the U.S. Geological Survey National Water Quality Network and comparisons to benchmarks for water years 2013–17.
Headnote: For the three types of carbon and for fluoride and boron, the counts of field blanks, detections in field blanks, and other statistics related to field blanks exclude detections that are attributable to nonfield sources of contamination, such as the source solution or a laboratory-derived source, that have no bearing on potential contamination of environmental samples and also exclude samples that were collected with a pankow sampler. ND, nondetection; —, not applicable.
Explanation of selected column names and contents
Compound group: N, nutrients; C, carbon; P, physical; S, sediment; M, major ion; T, trace element; Pe, pesticide or degradate
Unit: mg/L, milligram per liter; u/cm, unit per centimeter; µg/L, microgram per liter; ng/L, nanogram per liter; 
Num SW samples: number of surface-water samples
Num FB samples: number of field-blank samples
Num detections in FB: number of detections in field blanks
% FB detections: percent of field blanks with a detected result
FB 90%/95th: concentration of field blank that is the 90 percent upper-confidence limit for the 95th percentile of all field blanks
FB max: maximum concentration of field blank
SW min: minimum concentration of detected surface-water result
SW 1st%: first-percentile concentration of detected surface-water results
HHB concentration: concentration of human-health benchmark 
HHB type:  type of human-health benchmark: MCL, maximum contaminant level; HHBP–NC, EPA chronic noncancer human-health benchmark for pesticides; HHBP–C, EPA carcinogenic human-health benchmark for pesticides; HBSL–C, U.S. Geological Survey (USGS) cancer health-based screening level; HBSL–NC, USGS noncancer health-based screening level.
FB max fraction HHB: maximum concentration of field blank as a fraction of the given human-health benchmark
NC or ALB concentration: concentration of nutrient criterion or aquatic-life benchmark. The value presented in this column for pesticides is the lowest concentration if there are multiple types of aquatic-life benchmarks. 
NC or ALB type: type of nutrient criterion or aquatic-life benchmark. Min EPA, minimum U.S. Environmental Protection Agency (EPA) nutrient criterion for ecoregions; AF, acute fish; AI, acute invertebrate; ANVP, aquatic nonvascular plant; AVP, aquatic vascular plant; CF, chronic fish; CI, chronic invertebrate.
FB max fraction NC or ALB: maximum concentration of field blank as a fraction of the given nutrient criterion or aquatic-life benchmark.
FB 90%/95th fraction NC or ALB: concentration of field blank that is the 90-percent upper-confidence limit for the 95th percentile of all field blanks as a fraction of the given nutrient criterion or aquatic-life benchmark.
Max DL < lowest ALB?: True or False for whether the maximum detection limit is less than the lowest aquatic-life benchmark.
Max RL < lowest ALB?: True or False for whether the maximum reporting level is less than the lowest aquatic-life benchmark.
SW 1st% fraction HHB: first percentile concentration of detected surface-water results as a fraction of the human-health benchmark.
SW 1st% fraction ALB: first percentile concentration of detected surface-water results as a fraction of the aquatic-life benchmark.


Table 4. The 90-percent upper confidence limit for the 95th-percentile concentrations of nutrient, carbon, major ion, trace element, and select pesticides in field blanks of the U.S. Geological Survey National Water Quality Network and suggestions for censoring based on detections in field blanks for water years 2013–17.
Headnote: Contamination bias is indicated by field blanks. All forms of carbon and fluoride in table 3 field blanks show no evidence of potential for substantial contamination despite up to 95-percent detections in field blanks. For fluoride and boron, the count of field blanks, detections in field blanks, and other statistics related to field blanks exclude detections that are attributable to nonfield sources of contamination, such as the source solution or a laboratory-derived source, that have no bearing on potential contamination of environmental samples and also exclude samples that were collected with a pankow sampler. Some field-blank results for hydroxymetolachlor were excluded from the calculation of the 90%/95th concentration for reasons stated in the report pertaining to raised reporting levels. ND, nondetection; —, not applicable because there are too few samples for the calculation
Explanation of selected column names and contents
Compound group: N, nutrients; C, carbon; P, physical; S, sediment; M, major ion; T, trace element; Pe, pesticide or degradate
Unit: mg/L, milligram per liter; u/cm, unit per centimeter; µg/L, microgram per liter; ng/L, nanogram per liter
90%/95th in 2013, 2014, 2015, 2016, or 2017: 90-percent upper confidence limit for the 95th-percentile concentration of field blanks in the given water year.
Max RL: Maximum reporting level (RL) for water years 2013–17. Note this value does not necessarily reflect the RL in place in any particular year. For RL information by year, refer to table 1.
Contamination bias in 2013, 2014, 2015, 2016, or 2017: An "X" indicates potential contamination bias for the compound in the given water year.
Censoring suggested: RL, reporting level; 90%/95th, 90-percent upper confidence limit for the 95th-percentile concentration of field blanks. If only one or two field blanks are above the RL, censoring is suggested to the RL (for example, TKN in 2017). 


Table 5. Estimated variability in detection frequency of field replicates for samples in the U.S. Geological Survey National Water Quality Network for water years 2013–17.
Headnote: Pesticides are listed if there is at least one detected result in any of the pairs of replicate and environmental samples. The columns for the pairs with at least one detection are populated only for compounds that have at least 10 replicate pairs with a detection in 1 or both samples in the pair. Although the method code for particulate organic carbon (PARM_CD = 00689) changed from CAL06 to ALGOR in 2016, the result listed is a calculated value and the calculation (total particulate carbon minus particulate inorganic carbon) has not changed. NA, not applicable.
Explanation of selected column names and contents
Compound group: N, nutrients; C, carbon; P, physical; S, sediment; M, major ion; T, trace element; Pe, pesticide or degradate
Num all RP: Number of all replicate pairs
Num all RP both detections: Number of all replicate pairs where both results are detections
Num all RP both nondetections: Number of all replicate pairs where both results are nondetections
Num all RP inconsistent: Number of all replicate pairs where one result is a detection and the other result is a nondetection (i.e. inconsistent detections)
% all RP inconsistent: Percent of all replicate pairs where one result is a detection and the other result is a nondetection (i.e. inconsistent detections)
% >0detect RP inconsistent: For replicate pairs with at least one detection, percent that have inconsistent detections. This column excludes replicate pairs with no detections.
Mean detection rate >0detect: For replicate pairs with at least one detection, mean detection rate, in percent. This column excludes replicate pairs with no detections.
Variablity level: "Low" or "High" to indicate low or high variability in detection frequency (Martin, 2002; Bexfield and others, 2020b)


Table 6. Estimated variability in concentrations of field replicates for samples in the U.S. Geological Survey National Water Quality Network for water years 2013–17.
Headnote: Pesticides are listed if there are detections for both results of at least one replicate pair. —, not applicable; NA, the two-range model is not appropriate because there is no breakpoint in slope.
Explanation of selected column names and contents
Compound group: N, nutrients; C, carbon; P, physical; S, sediment; M, major ion; T, trace element; Pe, pesticide or degradate
Unit: mg/L, milligram per liter; u/cm, unit per centimeter; µg/L, microgram per liter; ng/L, nanogram per liter; NTRU, nephelometric turbidity ratio unit; µS/cm, microsiemens per centimeter at 25 degrees Celsius
Low concentration range: Low range of concentrations defined by the breakpoint in slope of the plot of mean concentration versus standard deviation of concentrations for replicate pairs; <, less than.
MSD low, break in slope: Mean standard deviation (concentration value) for points in the low concentration range
High concentration range: High range of concentrations defined by the breakpoint in slope of the plot of mean concentration versus standard deviation of concentrations for replicate pairs; >, greater than.
RSD high, break in slope: Mean relative standard deviation (in percent) for points in the high concentration range
MSD undivided: Mean standard deviation (concentration value) for undivided datasets where there is no breakpoint in slope for the plot of mean concentration versus standard deviation of concentrations for replicate pairs
Mean RSD, 2013, 2014, 2015, 2016, or 2017:  Mean relative standard deviation (in percent) in the given water year
High variability: Indicates year or years when variability is high, as determined by cases where mean relative standard deviation is greater than 20 percent for the indicated period of time


Table 7. Summary of results of data-quality assessment for inorganic compounds analyzed in surface-water samples of the U.S. Geological Survey National Water Quality Network based on all field quality-control sample types for water years 2013–17.
Headnote: —, no evidence of contamination bias or high variability; NA, not applicable.
Explanation of selected column names and contents
Compound group: N, nutrients; C, carbon; P, physical; S, sediment; M, major ion; T, trace element
Unit: mg/L, milligram per liter; u/cm, unit per centimeter; µg/L, microgram per liter
Benchmark concentration: concentration of human-health benchmark 
Benchmark type: Acute ALB, acute aquatic-life benchmark; MCL, maximum contaminant level; Min EPA, minimum U.S. Environmental Protection Agency (EPA) nutrient criterion for ecoregions; ; HBSL–NC, U.S. Geological Survey noncancer health-based screening level
Potential contamination bias in 2013, 2014, 2015, 2016, or 2017: An "X" indicates potential contamination bias for the compound in the given year as indicated by field blanks
High variability in 2013, 2014, 2015, 2016, or 2017: An "X" indicates potential high variability for the compound in the given year as indicated by field replicates reflecting variability in detection frequency or variability in concentration


Table 8. Summary statistics for pesticide field matrix spikes in samples of the U.S. Geological Survey National Water Quality Network for water years 2013–17.
Headnote: All statistics are for recovery values, in percent.
Explanation of selected column names and contents
RSD: relative standard deviation
Median recovery <70%: An "X" indicates median recovery is less than 70 percent; —, not applicable.
Median recovery >130 or F-pseudo >30%: An "X" indicates median recovery is less than 70 percent or F-pseudosigma is greater than 30%; —, not applicable.


Table 9. Summary of results of data-quality assessment for pesticides analyzed in samples of the U.S. Geological Survey National Water Quality Network based on all quality-control sample types for water years 2013–17.
Headnote: Information in this table on laboratory blanks (LBs), laboratory spikes (LSs), and third-party blind-spike samples for compounds not in the spike mixture (BSSs), and comments in the Notes column, is from Bexfield and others (2020b).
Explanation of selected column names and contents
Benchmark concentration: concentration of human-health or aquatic-life benchmark in nanograms per liter. The value presented is the lowest concentration if there are multiple types of benchmarks. 
Benchmark type: type of human-health or aquatic-life benchmark. MCL, maximum contaminant level; HHBP–NC, U.S. Environmental Protection Agency (EPA) chronic noncancer human-health benchmark for pesticides; HHBP–C, EPA carcinogenic human-health benchmark for pesticides; HBSL–C, U.S. Geological Survey (USGS) cancer health-based screening level; HBSL–NC, USGS noncancer health-based screening level; AF, acute fish; AI, acute invertebrate; ANVP, aquatic nonvascular plant; AVP, aquatic vascular plant; CF, chronic fish; CI, chronic invertebrate. The type presented in this column for pesticides corresponds to the lowest concentration if there are multiple types of aquatic-life benchmarks.
Evidence of potential contamination bias: Type of sample providing evidence of contamination bias. Possible sample types are LB, laboratory blank; BSS, third-party blind spike sample; FB, field blank.
Potential contamination bias in 2013, 2014, 2015, 2016, or 2017: An "X" indicates potential contamination bias in the given water year; —, not applicable.
Evidence of potential recovery bias: Type of sample providing evidence of substantial recovery bias and description of whether recovery bias is low or high. Possible sample types are LS, laboratory spike; BSS, third-party blind-spike sample; FMS, field matrix spike.
Potential recovery bias in 2013, 2014, 2015, 2016, or 2017: An "X" indicates potential recovery bias in the given water year; —, not applicable.
Evidence of potential overall high variability: Type of sample providing evidence of the potential for overall high variability. Possible sample types are LS, laboratory spike; BSS, third-party blind-spike sample; FRi, field replicates based on inconsistent detections; FRc, field replicates based on concentrations.
Potential high variablity in 2013, 2014, 2015, 2016, or 2017: An "X" indicates potential for overall high variability in the given water year; —, not applicable.
VQC: Validation quality code, —, no qualification, E, permanent estimated remark code (Sandstrom and others, 2015), E (2016), estimated remark code that went into effect January 1, 2016
Notes: 
   Degradation in blind spike might affect high variability: This compound has substantially higher variability in recovery for BSSs than for LSs, indicating that (1) recoveries for BSSs might be affected by potential degradation and (or) (2) the typically lower and more variable concentrations at which compounds are spiked in BSSs contributed to variability in BSS recoveries being outside acceptable limits.
   Degradation in blind spike might affect recovery bias: A parent compound with low recovery in BSSs or a degradate compound with high recovery in BSSs, but with good recovery (85 to 110 percent) in LS, indicates there is the potential for degradation to contribute to BSS recovery being outside acceptable limits, although the typically lower and more variable concentrations at which compounds are spiked in BSSs also could be a contributing factor. FMS recovery is also outside acceptable limits where "FMS" is also listed.
   Inconclusive evidence of contamination bias: Detection frequency in LBs and FBs is less than 1 percent, which does not provide clear support for the relatively high false-positive occurrence in BSSs. Also, with the exception of fenamiphos sulfoxide, false-positive results are not accompanied by indicators of interference. These observations might imply that spiking issues or postspiking degradation could be a contributing factor to false-positive results in BSSs.
   Matrix effect in field matrix spike might affect recovery bias: This compound has substantially higher or lower recovery in FMSs than in LSs and BSSs, indicating the potential for matrix effects to contribute to FMS recovery being outside acceptable limits.