POST-FIRE CHLOROPHYLL-A DYNAMICS IN HEADWATER RESERVOIRS AND A TOTAL ORGANIC CARBON DECISION SUPPORT SYSTEM FOR MUNICIPAL OPERATORS IN THE CACHE LA POUDRE WATERSHED

Abstract

Severe wildfires increasingly threaten western municipal water supplies by altering headwater catchments. This thesis addresses source water risks in the Cache la Poudre (CLP) River watershed following the 2020 Cameron Peak Fire through two separate studies: evaluating upstream reservoir responses post-fire and developing operational decision-support tools for municipal treatment operators to track Total Organic Carbon (TOC).First, I evaluated wildfire impacts across seven high-elevation reservoirs and their downstream reaches using field sampling and remote sensing. Burned reservoirs exhibited elevated sediment (18%) and potassium (11%) concentrations, and elevated spring nitrate levels beginning in the second-year post-fire. Extensively burned reservoirs displayed consistently higher total dissolved nitrogen, sulfate, and dissolved organic carbon. Chlorophyll a (Chl-a) concentrations were 18% higher in burned reservoirs, though remote sensing revealed no immediate secchi disk depth decrease amidst long-term warming trends. While mainstem Chl-a remained low (0.5–1.5 μg L⁻¹), reservoirs exerted localized influences on downstream Chl-a. These results indicate that reservoirs may modulate wildfire impacts by transforming nutrient enrichment into algal biomass that propagates downstream, but in this study downstream Chl-a remained low. Second, we designed two machine-learning tools to support municipal early warning of elevated in-stream TOC. To estimate upstream TOC concentrations in real-time, I trained an Extreme Gradient Boosting (XGBoost) model ensemble on multi-parameter sonde data. To support longer-term planning, I developed a generalized additive model (GAM) incorporating historical grab sample data (2010–2025), SNOTEL snowmelt rates, and NOAA streamflow forecasts to predict TOC seven days in advance at the CLP raw water intake. Both tools were integrated into a live operational dashboard. Developed through an iterative, stakeholder-driven process with municipal partners, this framework demonstrates how academic-practitioner partnerships can successfully transition environmental data science into operational water management.