Date of Award

Spring 5-9-2017

Document Type


Degree Name

Doctor of Philosophy (PhD)

First Advisor

JoAnn Silverstein

Second Advisor

Balaji Rajagopalan

Third Advisor

Rita Klees

Fourth Advisor

Jessica Kaminsky

Fifth Advisor

Paul Chinowsky


By 2050, it has been estimated that 70 percent of the world’s population will live in cities, concentrating waste as well as local environmental stresses. At the same time, decentralized approaches to sanitation are projected to grow due to the capital cost of sewers and centralized treatment facilities. Yet the common belief that technology will assure the performance of on-site systems over their life cycle may lead to significant underestimation of the actual risks to public and environmental health from owner-operated residential sanitation systems. Safe on-site storage, transformation and disposal of human waste require knowledge of how factors such as individual ownership, operations and management, and scale impact wastewater treatment reliability, risk and resilience under both normal and extreme conditions.

This dissertation research is developed to fill a gap in performance-based knowledge of OWTS function, especially the likelihood of system failure over lifetime operation. As such, a data-based investigation of highly decentralized and privatized wastewater management represented by on-site wastewater treatment systems (OWTS) was conducted using data from OWTS located in Boulder County, Colorado. Data were acquired from County maintained repair permit application records, inspection documentation, and property attributes. Methods are developed to quantitatively diagnose components that determine OWTS life cycle performance such as reliability, risk, fragility, and resilience by applying commonly used statistical modeling approaches based on the Generalized Linear Model regression method. Statistical modeling is then applied to analyze two conditions not controlled by current OWTS design and siting regulations: owner behavior and weather-related hazards.

Statistical model results confirm that owner-operations significantly affect life cycle OWTS functionality. Specifically, the results indicate the significant benefit of regulated inspections and maintenance as means to ensure that once installed, these systems continue to perform reliably and cost-effectively over their lifetime. Although a significant public information campaign (SepticSmart) has been maintained by the Boulder County Public Health Department, it is evident from the regression analysis that relying solely on public education about the importance of practices such as regular inspection and maintenance is insufficient to positively influence private owners’ decisions and prevent generation of externalities such as contaminant release from failed OWTS.

A resilience framework is developed to demonstrate the degree to which decentralization influences systematic OWTS vulnerability to weather – both wetter-than-average conditions and extreme storm events, independent of individual OWTS operations. Widespread natural hazards such as flooding are found to affect the frequency and degree to which OWTS function is lost, and more importantly delay their recovery, attributable in part, to a demand surge for both materials and repair services when multiple systems fail simultaneously. Longer recoveries are likely to have environmental and public health consequences due to the prolonged release of contaminants as well as secondary costs related to homeowner losses resulting from a failed OWTS.

Ultimately, the findings of this dissertation contribute to the decisions of planners, regulators and community stakeholders concerned with varying levels of wastewater treatment reliability, risk and resilience along the sanitation continuum from highly centralized and regulated collection and treatment infrastructure to relatively unregulated onsite systems operated by their owners. This research has demonstrated the importance of factors representing two heretofore unrecognized dimensions of OWTS life cycle performance – behavior of individual owners and enhanced vulnerability to natural hazards, and thus enables planners to decide if lar