Overview
Students will get hands-on experience with real-world water quality monitoring technology. The Virridy Lume sensor uses tryptophan-like fluorescence (TLF) — an optical method that detects microbial contamination without reagents or lab access. You will deploy the Lume in Boulder Creek, collect grab samples for laboratory analysis using IDEXX Colilert-18 (the EPA-approved reference method), and then compare the sensor readings to lab results using basic data analytics.
This lab connects directly to course themes: appropriate technology, sustainability engineering in practice, monitoring & evaluation, and the role of data in evidence-based solutions.
Learning Objectives
- Understand how optical sensors detect microbial contamination in water
- Practice field sampling protocols for water quality monitoring
- Learn the IDEXX Colilert-18 reference method for E. coli enumeration
- Compare continuous sensor data to discrete grab samples
- Apply basic statistical analysis to evaluate method agreement
- Critically assess the advantages and limitations of real-time vs. laboratory-based monitoring
Lab Structure
Session 1
Field Deployment
75 min • Boulder Creek
Field
Lab
18–22 hrs incubation
Session 2
Data Analysis
75 min • Computer Lab
Computer
Session 1: Field Deployment (75 min, Boulder Creek)
Location: Boulder Creek at the CU Boulder campus (SEEC building area)
Before Class
Required Preparation
- Read: Virridy Lume technology overview
- Read: ATP Comparability Report summary
- Watch: Introduction to water quality monitoring (5 min video — posted on Canvas)
In the Field
10 minutes
Introduction
Instructor overview of the Lume sensor — TLF principle (excitation 275nm, emission 340nm), what tryptophan indicates about microbial contamination, how the ML model converts fluorescence to E. coli estimates.
15 minutes
Sensor Deployment
Students deploy the Lume sensor in Boulder Creek. Observe proper submersion depth, orientation, and data transmission check via cellular.
30 minutes
Sample Collection
In groups of 3–4, collect 3 grab samples from Boulder Creek at the same location as the Lume:
- Use sterile 100mL Whirl-Pak bags
- Record time, GPS coordinates, water temperature, visual observations (color, turbidity, flow)
- Follow aseptic technique — no touching inside of bag, collect facing upstream
- Label samples with group number, time, location
- Samples must be processed within 6 hours (maintain on ice)
20 minutes
Colilert Processing Back in SEEC Lab
- Add Colilert-18 reagent to 100mL sample in Quanti-Tray
- Seal in Quanti-Tray/2000 using the sealer
- Place in incubator at 35°C for 18–22 hours
- Record sample IDs and timestamps
Safety Notes
- Wear nitrile gloves during sampling and lab work
- Do not ingest creek water
- Wash hands thoroughly after field work
- Report any cuts or open wounds before handling water samples
- Wading boots or waterproof footwear recommended
Session 2: Data Analysis (75 min, Computer Lab)
Before Class
Required Preparation
- Read Colilert-18 results (instructor will photograph Quanti-Trays and share)
- Download Lume data from the class dashboard
In the Lab
15 minutes
Read Colilert Results
- Count positive wells (yellow + fluorescent) on each Quanti-Tray
- Use the MPN table to convert well counts to MPN/100mL
- Record E. coli concentration for each sample
- Discuss: What does MPN mean? Why is it “Most Probable Number” and not an exact count?
10 minutes
Access Lume Data
- Open the Lume data dashboard
- Download CSV of continuous TLF, temperature, and turbidity readings
- Identify the time windows corresponding to your grab samples
40 minutes
Data Analysis Exercise
Using Google Sheets or Excel, complete Parts A–D below.
10 minutes
Class Discussion
Each group shares their key finding. Compare results across groups — did everyone get similar Colilert results? Discuss: why might grab samples from the same location at the same time differ?
Part A
Basic Comparison
- Create a table: Sample ID, Collection Time, Colilert MPN/100mL, Lume TLF reading at same timestamp, Lume E. coli estimate
- Calculate the difference and percent difference between methods
- Plot: Scatter plot of Lume estimate (x-axis) vs. Colilert MPN (y-axis)
Part B
Continuous vs. Discrete
- Plot the full Lume time series for the deployment period (TLF, temperature, turbidity)
- Mark your grab sample times on the plot
- What variability do you see in the continuous data that the grab samples miss?
- Calculate: mean, min, max, and standard deviation of Lume readings during the sampling period
Part C
Water Quality Classification
Apply the EPA recreational water quality criteria to classify each sample and each Lume reading:
| E. coli Concentration | Classification | Meaning |
|---|---|---|
| < 126 MPN/100mL | Safe | Safe for primary contact (swimming) |
| 126 – 410 MPN/100mL | Monitor | Elevated; continued monitoring recommended |
| > 410 MPN/100mL | Unsafe | Unsafe for primary contact |
- Do the Colilert and Lume classifications agree for each sample?
- Bonus: Calculate Cohen’s Kappa agreement statistic
Part D
Critical Analysis
- What are the advantages of continuous monitoring vs. grab sampling?
- What are the limitations of TLF as a proxy for E. coli?
- If you were designing a monitoring program for Boulder Creek, which method would you use and why?
- How does this connect to the carbon credit monitoring challenges discussed in class (Virridy’s dMRV)?
Deliverable
Lab Report — Individual, 3–5 pages
Submit as PDF via Canvas by the deadline posted on the course schedule.
- Methods: Describe sampling protocol, Colilert procedure, and Lume deployment
- Results: Include your comparison table, scatter plot, time series plot, and classification table
- Discussion: Address Part D questions. Connect to course themes (technology validation, sustainability monitoring, evidence-based engineering)
- Conclusion: Would you recommend the Lume for routine water quality monitoring in Boulder Creek? Why or why not?
Grading Rubric (100 points)
| Component | Points |
|---|---|
| Methods description | 15 |
| Data table and calculations | 20 |
| Scatter plot (Lume vs. Colilert) | 15 |
| Time series analysis | 15 |
| Classification analysis | 10 |
| Discussion (Part D) | 15 |
| Conclusion and recommendation | 10 |
| Total | 100 |
Reference Materials
Use these resources for background reading and your lab report: