Project Overview
This project reflects an integrated convergence framework combining continuous environmental sensing, machine learning analytics, regulatory co-design, and carbon-finance mechanisms to address water quality and climate challenges simultaneously. The approach links real-time microbial water quality monitoring with carbon credit generation from drinking water treatment, precision irrigation, and watershed restoration programs—creating financing mechanisms that fund environmental improvements while generating verified, tradeable emissions reductions. The Virridy Lume sensor platform provides continuous in-situ measurement of tryptophan-like fluorescence (TLF), chlorophyll-a, and fluorescent dissolved organic matter (FDOM), operating autonomously with up to one year of battery life, no regular calibration or cleaning, and cellular/satellite data reporting at intervals from 30 seconds to 24 hours. Field validation across 500+ coincident samples has demonstrated >94% categorical accuracy for E. coli estimation with site-specific calibration, 7% MAPE in log-transformed concentration space, and >90% balanced accuracy on the Seine River in Paris. The Lume uses a Silicon Photomultiplier (SiPM) detector—the only TLF sensor to do so—providing 12,000x greater responsivity than competing photodiode-based instruments, at $200/month versus $5,000–$30,000 for legacy instruments.
The project currently operates across nine countries—the United States, Rwanda, Burundi, Democratic Republic of Congo, Madagascar, Kenya, Tanzania, Turkey, and France—with programs spanning drinking water treatment, precision irrigation, and watershed restoration. Carbon credits are generated and verified under Gold Standard, Verra VCS (VM0042), and Regen Registry methodologies, with projections of over 3 million credits and 5 million people reached by 2030. See virridycarbon.com for program details and thelume.ai/research for sensor validation data.
Intellectual Merit
The project advances three distinct research frontiers. First, we developed the Lume—the only water quality sensor using Silicon Photomultiplier (SiPM) detection for tryptophan-like fluorescence—and validated it across 500+ coincident field samples achieving >94% categorical accuracy for E. coli estimation. Second, we built and are scaling a global carbon credit portfolio spanning drinking water treatment programs across six African countries (Rwanda, Kenya, Burundi, DRC, Madagascar, Tanzania), precision irrigation with Netafim in Turkey and Mexico under Verra VM0042, and watershed restoration in the United States—including the first-ever watershed carbon credit issuance (24,143 verified tCO2e through Regen Registry). Third, we enacted state legislation (Colorado SB24-037) that directs a regulatory agency to collaborate on water quality trading pilots incorporating climate finance—a first-of-its-kind policy-technology coupling.
The sensing component advances environmental monitoring science through development of autonomous fluorescence instrumentation—the first single-unit fluorimetric sensor for continuous microbial water quality monitoring—capable of measuring tryptophan-like fluorescence at 280/350 nm excitation/emission, chlorophyll-a at 470/680 nm, and FDOM at 365/480 nm, with integrated turbidity and temperature compensation. Under controlled laboratory conditions, the system reproducibly demonstrated sub-ppb sensitivity that exceeds the stated detection limits of many commercially available TLF instruments. These results are supported by a submitted manuscript (Knopp et al., 2026, EarthArXiv) and build on prior peer-reviewed work including Bedell et al. (2022, Water Research). The underlying sensing approach is protected by U.S. Patent 11,506,606 (Alarm Threshold Microbial Fluorimeter) and U.S. Patent 11,507,861 (Machine Learning Techniques for Improved Water Service Delivery), with five additional patent applications pending.
The Lume is differentiated from competing instruments across multiple dimensions. It employs a Silicon Photomultiplier (SiPM) detector—no other water quality TLF sensor uses this technology—providing 12,000 times greater responsivity and over 100,000 times greater internal gain than the photodiode detectors in all competing instruments, at a detector cost of approximately $130 compared to $4,000 for conventional PMTs. Beyond the detector, the Lume is the only TLF sensor that outputs quantified E. coli risk estimates rather than raw fluorescence units, using gradient-boosted machine learning models trained on 500+ coincident field samples. It requires no regular calibration or cleaning (competitors require site-specific calibration and frequent maintenance), operates for up to one year on battery (vs. wired power for most competitors), integrates cellular and satellite IoT connectivity with a cloud dashboard (competitors require third-party telemetry), and is priced at $200/month as a service versus $5,000–$30,000 one-time instrument purchases. The only direct competitor—Proteus Instruments (UK, ~$561K annual revenue)—sells a $30,000 multi-parameter sonde that is significantly larger, heavier, more complex to deploy, and achieves lower TLF sensitivity and E. coli estimation accuracy than the Lume. Proteus has already been displaced by the Lume on the H2NOW real-time water quality platform in Chicago, and on monitoring of the Marne River in France by the University of Paris.
Drinking Water
Binary classification at regulatory thresholds
Chlorinated supply monitoring
Natural Waters
Local model — Boulder Creek, Colorado
Global model — multi-site cross-validation & Seine River, Paris
Field deployments showed strong agreement between sensor-derived E. coli estimates and laboratory Colilert measurements, with categorical accuracy exceeding 94% under site-specific calibration and approximately 23% MAPE from a global model in log-transformed concentration space.
A second intellectual contribution lies in life-cycle accounting frameworks that quantify emissions avoided through infrastructure substitution. Attributional life-cycle assessment of a watershed adaptive management program demonstrated verified avoided emissions of 24,143 tCO2e for 2017–2022 and projected approximately 73,463 tCO2e net avoided emissions over 2017–2036 (Johnson et al., 2026).
A third contribution is the integration of policy design, behavioral science, and market mechanisms into environmental engineering research. Comparative analysis of water-quality trading programs across multiple states demonstrates that institutional clarity, enforceability, and financial viability are primary determinants of implementation success.
Broader Impacts
The project produces measurable environmental, societal, economic, and public-health benefits across multiple geographic contexts. Peer-reviewed randomized controlled trials of the project’s water treatment programs have demonstrated 29–49% reductions in childhood diarrhea, with a cost-benefit ratio of 5.6x across health, livelihood, and environmental dimensions.
Carbon-financed water programs across six African countries (Rwanda, Kenya, Burundi, DRC, Madagascar, Tanzania) currently serve over 2.6 million people across 1,180 monitored sites, with projections exceeding 5 million by 2030. In Turkey, precision irrigation programs cover over 1,000 hectares, reducing emissions by 3.5 tCO2e/hectare annually. In Wisconsin, the Yahara WINS program across 139,000 hectares produced the first-ever verified watershed carbon credit issuance with over 73,000 tonnes CO2e avoided.
The project produces large-scale scientific infrastructure including a global E. coli database exceeding 10 million bacterial coliform observations and 13 peer-reviewed publications since 2023 in journals including Nature Communications, Science of the Total Environment, Water Research, and ES&T Water. Two granted U.S. patents (11,506,606 — Alarm Threshold Microbial Fluorimeter; 11,507,861 — Machine Learning Techniques for Improved Water Service Delivery) and five pending patent applications protect the core intellectual property.
The Lume’s economic model also advances accessibility. Conventional grab-sample testing using IDEXX Colilert costs approximately $19 in consumables plus technician labor and vehicle costs per sample, with 18–24 hour time to result. The Lume provides 720 readings per month per site at $200/month—a 180-fold increase in temporal resolution at comparable or lower cost, while capturing transient contamination events that weekly sampling misses entirely.
Team
The project team is deliberately structured to reflect a convergence model that integrates expertise from multiple disciplines: instrumentation engineering, machine learning, environmental chemistry, hydrology, watershed science, behavioral science, economics, public policy, and climate finance. This multidisciplinary configuration ensures that scientific discovery, technology development, policy implementation, and market deployment proceed simultaneously rather than sequentially. Leadership is provided by Principal Investigator Evan Thomas, PhD, CEO and founder of Virridy and Director and Endowed Chair of the Mortenson Center in Global Engineering and Resilience at the University of Colorado Boulder (PhD Aerospace Engineering, MPH, MBA; former NASA civil servant). The executive team includes Danny Wilson, PhD (CTO; PhD Mechanical Engineering, UC Berkeley; Fulbright and NSF Fellow), Alex Johnson (Chief Strategy Officer; former VP at The Freshwater Trust; $40M+ in water quality trading credits sold), and Laura MacDonald, PhD (International Carbon Programs Director; PhD Geography and Environmental Engineering, Johns Hopkins University). Co-Investigators include Matt Ross, PhD (Associate Professor of watershed science, Colorado State University; PhD Ecology, Duke), Krister Andersson, PhD (Professor of Political Science, University of Notre Dame; behavioral science and policy design), and Carlo Salvinelli, PhD (Associate Teaching Professor, CU Boulder; environmental engineering and water treatment systems). Research staff include Whitney Knopp (Environmental Engineer; PhD candidate, CU Boulder; lead author on the Lume sensor validation manuscript) and John Ecklu (Environmental Engineer; East Africa implementation and carbon market applications).
| Sector | Partner | Role |
|---|---|---|
| Academic | University of Colorado Boulder, Colorado State University (Open Current spin-out) | Research + modeling |
| State Government | CDPHE, EPA Region 8 | Regulatory integration |
| Federal | NSF, NASA, U.S. Air Force | Funding + testing |
| Technology | In-Situ Inc., Urban Sky, Blues | Hardware + connectivity |
| Nonprofit | Friends of the Yampa, MWA | Community implementation |
| Carbon buyers | Louis Dreyfus Company, PetroChina, Water Environment Federation, Mortenson Construction | Credit purchases |
| International | Netafim, BGS, City of Paris | Programs + pilots |
| Utilities | Boulder, Denver, Madison + others | Pilot implementation |
Intellectual property includes two granted U.S. patents—11,506,606 (Alarm Threshold Microbial Fluorimeter) and 11,507,861 (Machine Learning Techniques for Improved Water Service Delivery)—and five pending patent applications covering DMRV fusion networks, advanced sensing modalities, and related technologies.
Project Accomplishments to Date and Plan for Phase 2, Year 2
Across Phase 2 Year 1, progress has continued along all nine objectives. On the sensing and technology front (Objectives 6–8), the team built and tested the Lume v1.1 sensor hardware, collected more than 500 coincident field samples comparing sensors to lab E. coli measurements across eight installations, and developed machine-learning models achieving 70–80% categorical accuracy and 15–30% log error rates. Laboratory testing demonstrated consistent TLF detection through turbidity of 0–500 NTU. The global E. coli dataset has reached over 10 million observations with more than 8 million passing quality checks. A manuscript has been submitted reporting sub-ppb sensitivity and greater than 90% balanced accuracy on the Seine deployment.
On the carbon finance front (Objectives 1–3), the project has signed credit sales contracts totaling approximately $17M (minimum, with additional upside sharing and option agreements) in projected revenue. Actual verified credit issuances to date total 78,103 tCO2e: Amazi Meza Rwanda (7,665 in 2024 and 33,911 in 2025), LifeStraw Kenya (12,384 in 2026), and Yahara WINS Wisconsin (24,143 in 2026). Projected issuances across the full nine-project portfolio—spanning Rwanda, Kenya, DRC, Burundi, Madagascar, Tanzania, Turkey, and Wisconsin—are expected to exceed 680,000 credits annually by 2030. Key milestones completed include Rwanda Letter of Authorization, Turkey PDD listing under Verra, first watershed carbon credit issuance under Regen Registry, and the Water Environment Federation’s purchase of 1,000 tonnes to offset WEFTEC. The team finalized approximately 100,000 credits from Water Mission in Tanzania. Louis Dreyfus Company has contracted interest in irrigation-based credits through VM0042.
On deployment and market traction, the Lume sensor network is expanding rapidly across domestic and international sites. Active deployments and partnerships include: Suez Water and the City of Paris on the Seine and Marne Rivers; the British Geological Survey on the Thames River (UK); EPA and San Diego State University at Imperial Beach, California; the Indian River Lagoon with NASA Kennedy Space Center and Marine Resources Council (Florida); Current on the Chicago River; the Charles River Watershed Association in Boston; the Bow River Watershed Council in Banff, Canada; the Millennium Water Alliance in Kenya; and Virridy’s own Amazi Meza program in Rwanda. City of Boulder Utilities is a paying Lume customer, with additional contracted deployments underway with Denver Water and the City of Denver, Manchester Bay (Manchester by the Sea, MA), and Fort Myers, FL. The U.S. Air Force has confirmed intent to purchase 1–2 Lume v1.2 devices, with USAF personnel visiting the Virridy lab in April 2026. Channel partner discussions continue with Suez (Veolia) and OTT Hydromet.
The project’s cross-track integration focus is a collaboration with Daniel Yeh’s team and colleagues at NASA Kennedy Space Center, using Virridy’s Lume sensors to track algae-based water treatment performance both pre- and post-processing. This work leverages the Lume’s chlorophyll-a (470/680 nm) and TLF (280/350 nm) sensing modes to provide continuous, real-time monitoring of treatment efficacy in systems that use algae for nutrient removal and water reclamation—a direct convergence between the project’s sensing technology and NASA’s water treatment research.
On the policy front (Objectives 4, 5, 9), the project team advanced state legislation SB24-037, which was enacted in Colorado enabling watershed pilot programs. CDPHE approved three regional pilots and a three-year implementation timeline. The comparative analysis of national water quality trading best practices is 95% complete. One discharger is advancing toward pilot implementation. National engagement continues through the Aspen National Water Strategy and the Walton-backed Resilience Accelerator.
| Risk | Status | Mitigation |
|---|---|---|
| Manufacturing delays/tariffs | Resolved | Lume v1.2 received and in testing |
| Cl-A/FDOM market competition | Monitoring | Focus on TLF microbial-risk product; evaluating adjacent sensing modalities |
| USDA RCPP not awarded | Pivoted | Direct implementation via Netafim Turkey/Mexico |
| Partner sensor data tracking | In progress | Structured tracking processes being designed |
| Offtake contract coverage | Secured | All offtake contracted through 2031 |
| Metric | Target | Actual / Current | Status |
|---|---|---|---|
| Sensor categorical accuracy | 90% | >94% (site calibrated) | Exceeded |
| Coincident field validations | 500 | 500+ across 8 sites | Met |
| Global E. coli dataset | 10M observations | 10M+ (8M+ QC passed) | Met |
| Watershed carbon credits | First issuance | 24,143 tCO2e issued | Met |
| Carbon credit contracts | $10M | $17M+ executed | Exceeded |
| Countries operating in | 9 | 9 | Met |
| Regulatory pilots approved | 3 | 3 (CDPHE) | Met |
| Peer-reviewed publications (since 2023) | 10 | 13 | Exceeded |
| US deployment sites | 10 | 16+ active/planned | Exceeded |
The project’s nine objectives were originally scoped across a three-year Phase 2 timeline. As the metrics table above demonstrates, the majority of milestones have been met or exceeded in Year 1 alone—including first watershed carbon credit issuance, $18M in executed sales contracts, >94% sensor accuracy, enacted legislation, and 16+ deployment sites. This accelerated progress allows Year 2 to shift emphasis from proving feasibility to scaling validated systems and translating research outputs into operational, revenue-generating products and programs.
Year 2 will focus on three acceleration tracks. On the sensing side, the priority shifts from sensor validation to commercial deployment at scale: expanding the Lume v1.2 network from pilot sites to contracted municipal and utility customers, building site-generalizable E. coli models from the growing multi-site dataset, operationalizing partner-facing workflows for data access and device management. The USAF procurement (1–2 units for Wright-Patterson AFB testing, April 2026 lab visit) and Veralto strategic partnership negotiations represent pathways to institutional and channel-partner scale that were not anticipated this early in the program.
On the carbon finance side, Year 2 shifts from first issuances to portfolio-scale execution: completing the Water Mission Tanzania credit purchase, scaling VM0042 farmer recruitment in Turkey and Mexico to over 10,000 hectares (10x the current Konya Basin program), and closing the $20M investment fund with Total Impact Capital ($12M in verbal commitments secured, led by Bridges Outcomes Partnerships). With all offtake contracted through 2031, the focus is on supply-side scaling—generating credits fast enough to meet contracted demand.
On the policy side, Year 2 transitions from framework development to operational pilots: moving the SB24-037 Colorado pilots from “pilots in concept” to “pilots in motion” with at least one discharger advancing through CDPHE review, publishing the comparative national WQT analysis and the SB24-037 journal article as replicable frameworks for other states, and deepening engagement through the Aspen National Water Strategy and Walton Resilience Accelerator to position the project’s compliance model for national adoption.
Budget and Financial Overview
The NSF Convergence Accelerator award is structured as a deliverables-based contract, under which Virridy invoices $185,000 per month against monthly progress reports documenting milestone completion across all nine objectives. As of the report date, the team has submitted and invoiced a total of 9 monthly invoices for a cumulative total of $1.665M. This structure directly ties federal funding to demonstrated progress in sensing, carbon finance, and policy deliverables.
Major subcontracts support two key academic and implementation partners. The University of Colorado Boulder is subcontracted at $200,000 in Year 1, $200,000 in Year 2, and $100,000 in Year 3, providing research expertise in environmental engineering, watershed science, and sensor validation. Open Current (an NSF Engine recipient operating the H2NOW real-time water quality platform in Chicago) is subcontracted under the same structure ($200,000 Year 1, $200,000 Year 2, $100,000 Year 3), supporting deployment operations, partner engagement, and implementation activities. The remaining budget supports Virridy personnel, hardware development and manufacturing, field operations, carbon program development, and policy engagement.
The financial structure is complemented by non-NSF capital including private investment ($8.5M equity led by Accord Capital and Mortenson Construction), advance credit pre-purchases ($2M from Mortenson), and philanthropic grants (Moore Foundation, Walton Family Foundation). From a commercialization standpoint, the project is now coupled to revenue execution pathways including executed multi-year credit sales contracts totaling approximately $17M (minimum, with additional upside sharing and option agreements) in gross revenue, initial watershed carbon credit issuance through Regen Registry, and expanding Lume deployment contracts. Budget reallocations for Year 2 are primarily associated with scaling deployments, expanding MRV capacity, and selectively investing in sensor roadmap adjustments.
Sustainability Plan
Long-term sustainability is based on a hybrid research–commercialization model. The lead organization operates as a for-profit entity with dual revenue engines: water-quality sensing and analytics services (Lume deployments at $200/month per site) and carbon-credit generation and sales under Gold Standard, Verra VCS, and Regen Registry methodologies. Near-term sustainability is strengthened by executed credit sales contracts totaling approximately $17M (minimum, with additional upside sharing and option agreements), expanding Lume deployment contracts, USAF procurement interest, and active carbon revenue generation—the project is forecasting at least $3M in earned carbon credit revenue in 2026. On the sensor side, Virridy has produced 200 commercial Lume units and sold 60 year-to-date, with $557K in confirmed Lume revenue in 2026 YTD. The global water-quality sensor market was valued at $5.7 billion in 2024 and is projected to reach $9.1 billion by 2030 (8.1% CAGR).
The broader water quality sensor market is experiencing significant consolidation, with In-Situ Inc. acquired by Veralto Corporation for $435 million and Chelsea Technologies’ parent Covelya Group being acquired by Kraken Robotics for $615 million. Virridy is negotiating a strategic partnership with Veralto, which also owns Hach and OTT HydroMet, representing a potential channel to global distribution. Virridy has also received a $20M verbal acquisition offer from Geospace Technologies (Houston), further validating the commercial value of the platform. The project’s $20M carbon credit investment fund under development with Total Impact Capital has secured $12M in verbal commitments, led by Bridges Outcomes Partnerships.
Scalability derives from three reinforcing mechanisms: policy frameworks create regulatory pathways for watershed programs, monitoring technologies provide credible measurement and verification of outcomes, and carbon markets supply financing incentives. Together these create a self-reinforcing system enabling sustained expansion without reliance on continued grant funding. The project’s current trajectory—operating across 9 countries with projections of 5 million people reached and over 3 million credits generated by 2030—demonstrates that this model is already scaling beyond the initial research context.
Convergence Team Publications Cited
1. Knopp, W., Klaus, J., Wilson, D., et al. (2026). Advancing continuous in-situ quantification of microbial contamination in environmental waters using tryptophan-like fluorescence—Sensor design and validation. EarthArXiv preprint.
2. Demaree, K., Fankhauser, K., Cole, G., Ross, M., Thomas, E. (2026). Sensor Informed Predictive Model for Total Organic Carbon and Nutrients on the Upper Yampa River. ES&T Water.
3. Bedell, E., Harmon, O., Fankhauser, K., Shivers, A., Thomas, E. (2022). A continuous, in-situ, near-time fluorescence sensor coupled with a machine learning model for detection of fecal contamination risk in drinking water. Water Research.
4. Bedell, E., Sharpe, T., Purvis, T., Brown, J., Thomas, E. (2020). Demonstration of Tryptophan-Like Fluorescence Sensor Concepts for Fecal Exposure Detection in Drinking Water. Sustainability, 12, 3768.
5. Johnson, A., Cole, G., Quinn, J., Thomas, E. (2026). Avoided Greenhouse Gas Emissions from Watershed-Scale Adaptive Management: Life Cycle Assessment of Infrastructure Substitution in the Yahara WINS Watershed Carbon Project. Under review.
6. Landon, T., Johnson, A., Thomas, E. (2026). Feasibility of Green Infrastructure in Alternative Compliance Programs: Applying Water Quality Trading Experience to Colorado Senate Bill 24-037. Under review.
7. Aydın, A., Küroğlu, F., Thomas, E.A., Salvinelli, C., Polat, E.P., Yıldırak, K. (2026). Carbon Farming in Türkiye: Challenges, Opportunities and Implementation Mechanism. Sustainability, 18(2), 891.
8. Demaree, K., Johnson, A., Andersson, K., Thomas, E. (2025). The Nature-Based Paradigm Shift: Identifying and Overcoming Barriers to Sustainable Water Quality Solutions. ACS ES&T Water, 5(6), 2945–2952.
9. Ecklu, J., Thomas, E. (2025). Digital monitoring, reporting, and verification technologies supporting carbon credit-generating water security programs: State of the art and technology roadmap. Environmental Science & Technology Letters, 12(3), 251–260.
10. Demaree, K., Kurli, V., Magnuszewski, P., Andersson, K., Thomas, E. (2024). Development and evaluation of a digital behavioral economics game towards improved understanding of groundwater conservation in southern Colorado. PLOS Water, 3(12), e0000298.
11. Thomas, E. (2024). Turning global water security research into policy and action. PLOS Water, 3(7), e0000261.
12. Ecklu, J., Barstow, C., MacDonald, L., Fankhauser, K., Johnson, A., Adler, I., Ehrhardt, D., Thomas, E. (2024). Decarbonizing Water: The Potential to Apply the Voluntary Carbon Market toward Global Water Security. ACS ES&T Water, 4(6), 2655–2666.
13. Limb, B.J., Quinn, J.C., Johnson, A., Sowby, R.B., Thomas, E. (2024). The potential of carbon markets to accelerate green infrastructure based water quality trading. Communications Earth & Environment, 5(1), 185.