Methods, Benchmarks and Standards

The KICP water quality database is a comprehensive collection of data from multiple municipal entities, meticulously compiled to ensure it is comparable. Processing data for this comparability ensures that data from different sources can be meaningfully analyzed together, leading to more accurate, reliable, and actionable insights. The methods for monitoring, sample collection, and analytical analysis are detailed in the Boulder Creek and St. Vrain Coordinated Watershed Monitoring Framework (PDF).

Navigation Tip: Jump to the sections below using this linked list

• Background
• Preprocessing
• QA/QC
• Visualizations (Types of Graphs)
• Mann-Kendall Trends
• Benchmarks
• Standards
• Assessment

The data analysis evaluated changes in water quality over time (Temporal Analysis) and across different locations (Geographical Analysis). This effort examined monitoring sites from upstream-to-downstream and included various statistical metrics, such as annual averages (geometric means for E. coli), 25th and 75th percentiles, and comparisons between the study year and historical periods (5-year, 10-year, and full records). Additionally, the analysis assessed monthly aggregated values to explore seasonal variability within each basin, comparing the study year to historical data trends. More details on how these two types of analysis were applied are described below.

Measured values were also compared to current stream standards or draft criteria, which are established statewide by Colorado’s Water Quality Control Commission (WQCC) in the Basic Standards and Methodologies for Surface Water (Regulation No. 31 (PDF)) and are applied by each river basin. The Boulder and St. Vrain Watershed fall under Regulation No 38 (PDF), which provides water quality standards by each waterbody segment for the South Platte Basin.

Over the next decade, the State plans to update and develop additional water quality standards including nitrogen, phosphorus, chlorophyll a, and some metals. The WQCC will adopt and implement the new water quality criteria into Regulation No. 31 and prioritize implementation in basin regulations such as Regulation No. 38.

In addition, Regulation No. 85 (PDF) outlines the nutrient effluent limitations of domestic wastewater dischargers and management requirements for nutrient control.

Temporal Analysis

Changes over time were assessed by comparing monthly 2023 data to monthly data monitored over the previous 5-year period (2018 - 2022). Monthly means for each month in 2023, aggregated by basin, are compared to the same monthly mean calculations covering the 5-year period. These comparisons shed light on seasonal variability within each basin and on how 2023 conditions compare to a recent historical period. Specific monthly comparisons are included in graphs throughout the report to help illustrate the analysis, and monthly comparisons for all parameters can be found in the Explore tool. In the Explore tool, other historical periods may also be chosen for comparison, including the 10-year period from 2013 - 2022 and the full period of record available.

Geographical Analysis

Monitoring results were also compared along each major stream reach from upstream to downstream. Each location was assigned a river-mile as measured along the stream from the uppermost location (read more detail about this below in Visualizations - Upstream/Downstream Graphs).

These upstream-to-downstream assessments compare the annual mean values for each parameter at each location. This assessment provides insight into where along the streams the various constituents may be entering the waterway. Wastewater Treatment Facility discharge locations were included in these comparisons as well, which helps to understand how these inflows impacts or do not impact the watershed. Annual line charts as well as boxplots were developed to visualize the upstream-to-downstream data and are included throughout the report to help illustrate the analysis. They can also be reviewed for all parameters in the Explore tool.

Water Quality Scorecard

The water quality scorecard uses a Mann-Kendall test to determine statistically significant temporal trends in parameter concentrations since 2018. The results of this analysis have been combined into a single visualization, the Water Quality Scorecard. The scorecard is a table of parameters (as columns) and locations (as rows). Locations are ordered upstream-to-downstream and grouped by basin. In each cell of the table, trends are illustrated as arrow symbols. Annual mean concentrations are displayed in each cell and also illustrated with light (lower concentrations) to dark (higher concentrations) shading. This provides a picture of how parameter concentrations change in the streams moving upstream-to-downstream and also how concentrations have changed over the past six years (2018 - 2023).

While this effort focused on analysis of the 2023 data, to accurately compare to historical periods the full set of KICP’s historical data was also processed and reviewed. Steps taken to prepare data for the analyses and presentation included:

• Data covering the full period of record through 2023 for all locations and parameters that have been collected by the Partnership were imported into a centralized SQL database.

• Partner samples designated as field duplicates are used for QA/QC purposes only and were removed for this analysis.

• Values below detection limits and flagged as "ND", "BDL", "U" or "<" have been set to 1/2 the minimum detection limit (MDL) before averaging across time or geography.

• E. coli values below the minimum detection limit were set to 1 and values over the maximum detection limit were set to 2419.6 CFU/100ml.

• Values flagged as "J" (indicating estimated) were removed if a non-estimated value was also available for that sample.

• Different units for the same constituent were converted as necessary to align across the period of record.

Data Handling

Data from different sources varied in format, units, and methods of measurement, and a crucial step in the process involved normalizing the data to create a consistent and cohesive dataset. This process, often referred to as normalization, included converting all measurements to standardized units, aligning data collection methods to ensure comparability, and addressing discrepancies in data formats.

Additionally, GPS coordinates were used to match locations, ensuring that data from different sources corresponds accurately to the same geographical points. After undergoing rigorous quality assurance and quality control (QA/QC) procedures, the data has been systematically organized by site and date, allowing for seamless analysis and comparison across different locations, parameters and periods for a variety of different kinds of analyses.

Field Activities

Data providers are recommended to collect field blanks and field duplicates as 5% of total samples collected. If a field blank exceeds the reporting limit, the data provider should flag associated data points and eliminate the source of contamination. If a field duplicate exceeds a 25% relative percent different with the paired sample, the sample should be flagged, re-analyzed if possible, and the cause should be investigated.

Upstream/Downstream Graphs

Upstream-to-downstream line charts display monitoring locations along the mainstem of each main branch (St. Vrain Creek, Boulder Creek, and Coal Creek), as well as effluent discharge locations and the downstream-most monitoring location on key tributaries. Key characteristics of these graphs include:

• The X axis represents stream distance between monitoring points, with the uppermost point on each main branch as mile 0.

• The Y axis measures annual mean concentrations (or geomean for E. coli) for each location.

• Each main stream (Boulder, Coal and St. Vrain Creek) is shown as a separate series of connected points.

• Effluent locations are plotted according to their respective stream-mile location on each branch but shown as unconnected points with a different symbol.

• Tributary locations are also plotted according to the river mile location where that tributary joins the main branch, and are shown as unconnected points with a different symbol.

When interacting with the upstream-to-downstream graphs, hovering over a point will display the name of the monitoring location, its stream mile, and the annual mean value for the selected parameter.

Stream miles were determined by KICP staff using GIS. The uppermost location on each main branch was assigned a stream mile of 0, and distances were measured from there moving downstream following the bends of the creeks. The order of monitoring locations by main branch as

Example:

Monitoring Locations by Stream Mile

SEE STATION LOCATIONS ON A MAP: Monitoring Program - Where do we Monitor?

Boxplots

Boxplots are a visual tool used to summarize and compare the annual distribution of data, while providing a benchmark to assess trends over a longer period. The boxplot visualization graphs allow for a clear comparison of (2023) annual data distribution and 2023 median values, with the previous 5-year period (2018 - 2022) median providing a useful reference for detecting shifts or trends over time. Boxplots can vary in how they are configured. For this analysis, the Boxplots charts are set up as described and illustrated below.

The boxplots in this report include:

• Median (Central Tendency): The line inside the box represents the median, which is the middle value of the data, dividing all measurements it into two equal halves by count of measurements so that 50% of measurements are greater than the median and 50% are less than the median by value. The Median is eqivalent to the 50th percentile.

• Inter-quartile Range (IQR): The box itself shows the IQR, which spans from the 25th percentile (Q1) to the 75th percentile (Q3). This range captures the middle 50% of the data.

• Whiskers: The lines extending from the box, called whiskers, represent the range of data within 1.5 times the IQR from the quartiles. Due to limitations of the online graphing libraries, rather than representing the actual 1.5 x inner quartile values by station, the whiskers on these graphs have been set to the average of inner-quartile ranges for all stations within the basin, for the selected parameter and year/comparison period. A future update to this report will include refining the whiskers to reflect site-specific IQR values.

Example:

Monthly Average Line Graphs

These line graphs represent data that depict the monthly average for a parameter, in each stream for the study year, compared to the monthly average for the chosen period of record. Please note that interactive graphs throughout this report and in the Explore Section (default to a 5-year historical study period, but this may be changed to reflect a 10-year period or the full record available).

Example:

Trends for the 6-year period from 2018 - 2023 were calculated using a two-tailed Mann-Kendall test, which calculates the statistical significance of an upward or downward trend over time. It also determines a coefficient of variation (CV) which evaluates the variability of results over the study period. For this analysis, Mann-Kendall results have been characterized as follows:

• A statistical significance less than 90% indicates no significant change and is shown as Stable if the calculated CV is less than 1 and No Trend if the CV is greater than 1.

• Increasing or Decreasing trends are indicated by a 95% or higher significance associated with the trend.

• Probably Increasing or Probably Decreasing indicates there is an 90%-95% statistical significance associated with the trend.

Water quality measurements require context to be meaningful. To provide necessary context and allow for relative comparisons across the watershed, Benchmark values for key parameters have been established. Some benchmarks represent current or proposed regulatory standards while others were derived from scientific studies. The benchmarks are not intended to assess regulatory compliance but rather provide context for results.

Benchmarks were developed by review of relevant numeric stream standards for each stream, where available. In some cases stream standards are based on calculated values that vary with stream hardness, pH or temperature.

In the case of the calculated table value standards (TVS) based on hardness concentrations, an annual mean hardness value based on 2023 monitoring for each location was used to calculate the standard for the sake of comparison (these parameters are noted below). Cold Water standards applied to monitoring sites BC-Can, BC-CU and SBC-3.5. All other sites are in stream reaches classified as Warm water streams.

The annual mean pH value for the location.

The annual mean water temperature for the location in degrees Celsius.

The annual mean hardness for the location.

0.0577 / (1 + 10(7.688 - pH)) + 2.487 / (1 + 10(pH - 7.688)) x MIN(2.85, 1.45 x 10^(0.028 x (25 - TempC)) )

0.0577 / (1 + 10(7.688 - pH)) + 2.487 / (1 + 10(pH - 7.688)) x 1.45 x 10(0.028 - (25 - MAX(TempC, 7)))

e(0.8545 x ln(Hardness) - 1.7428)

e(0.3331 x ln(Hardness) + 5.8743)

e(1.72 x ln(Hardness) - 9.06)

0.986 x e(0.9094 x ln(Hardness) + 0.6235)

The Colorado Water Quality Control Commission adopts water quality classifications, standards and implementation processes for surface waters of the state, into Regulation #s 31 - 38. The Boulder, St. Vrain and Coal Creek watersheds are covered under Regulation No. 38.

A standards assessment was not conducted as part of this analysis, other than integration of applicable standards-based benchmarks, described in the previous section. However, some highlights of relevant standards affecting the Boulder, St. Vrain and Coal Creek watersheds are included here to provide context for the 2023 data.

Temperature

Temperature standards change based on the time of year and the classification of streams into warm and cold water tiers, reflecting expected fish species. Stream standards are assessed using continuous temperature data, therefore these point in time measurements are not compared to the state standard in this report. Temperature tiers, monitoring sites, and applicable standards are listed in the table below.

Dissolved Oxygen

Regulation 38 standards require cold water streams to maintain DO concentrations above 6 mg/L (≥ 7 mg/L during spawning) and warm water streams to remain above 5.0 mg/L to support aquatic life.

Nutrients

In Colorado, nutrient water quality standards focus primarily on regulating levels of nitrogen and phosphorus in surface waters. These nutrients are essential for aquatic life but can cause harmful algal blooms and degrade water quality when present in excessive amounts. The standards aim to protect aquatic ecosystems, drinking water supplies, and recreational waters. Currently, there are only interim criteria for these standards (except Ammonia) until new standards are adopted. The interim stream water quality nutrient standards are outlined in the table below.

Ammonia

Ammonia standards depend on pH, temperature, and stream classification. Generally, as pH and temperature increase ammonia concentrations can be more toxic to fish so the standard is lower and cold-water species, such as trout, are more sensitive than warm-water species. The equations for calculating the ammonia standard can be found in Regulation 31 (Table 2).

Key guidelines include:

• A single sample can only exceed the acute standard once every 3 years.

• The 85th percentile of samples must not exceed the chronic standard, based on the average pH and temperature.

• For individual events, no more than 15% of values should exceed the paired standard based on temperature and pH.

E. coli

Colorado assesses the recreational E. coli standard using a static two-month geometric mean with at least five samples. If the geometric mean exceeds the recreational standard of 126 CFU/100mL then the segment will be classified as impaired and placed on the 303(d) list.

Metals

Many water quality standards for metals are based on hardness due to its effect on metal toxicity. Hardness based standards can be calculated using a mean hardness or paired with a specific value that was analyzed from the same sample. These standards also consider other factors, including specific aquatic classifications.

Some metals, such as arsenic, have numeric standards that vary based on aquatic life and water supply classifications, rather than hardness. All metals standards, including the equations for the calculation of the table value standards can be found in table III of Reg. 31.

Key guidelines include:

• individual samples must not exceed the acute standard more than once every three years.

• The 85th percentile must not exceed the chronic standard.

Stream standards in Colorado are assessed using the 303(d) Listing Methodology as required by the Federal Clean Water Act. Assessment is completed using the Colorado Department of Public Health and Environment (CDPHE) current regulations and listing methodology.

Current listings for impaired waters can be found in Regulation No 93. or the Regulation #93 Dashboard and a detailed list of current use attainment can be found in the 2024 Integrated Use Attainment Table for Rivers.