UC Berkeley

Today, a billion people lack access to safe drinking water. Climate change, a host of

unsustainable practices, and emerging contaminants threaten to exacerbate the problem.

For the developing-world communities disproportionately affected, limited infrastructure and

resources often preclude the developed-world solutions. New technologies will be necessary.

Unfortunately, even mature technologies often fail in these settings. This work presents two

efforts: one to advance understanding of robust technical interventions and one to improve

water treatment technology for such settings.

In the Poshiir River watershed study, we built comparative case studies of single-village

piped-water drinking water schemes in rural Maharashtra. In-depth datasets were built from

direct technical observations, evaluation of records at national, district, and local levels, and

structured interviews with scheme users and administrators. Redundant information sources

were used to triangulate and assess confidence in our factual findings, which were formalized

in novel node-network models of each scheme-village pair. Hypotheses were developed and

mapped onto these networks via process tracing. We theorize that resilience of a scheme

is dependent on healthy, positive feedback loops within these social-technical networks. We

also took early steps toward typological theories on scheme failures.

In the course of the Poshiir study, we developed a method for rigorously and objectively

reconciling data from disparate sources. For a number of queries on a given topic, the

method takes the array of agreements and disagreements between all sources as inputs; it

generates most-probable values for the validity of each source's response, as well as mean

validities for the sources themselves. Early tests of the method with a verifiable dataset

provide meaningful results, and application to the Poshiir study have returned consistent

scores when assessing comparable cases. While further investigation is certainly required,

the method appears promising.

We also sought to improve the effective flow rate for capacitive deionization (CDI), a

treatment technology for removing ionic species from brackish water. We first proposed two

novel powering arrangements that would accomplish this by increasing effective ion mobility:

optimal powering profiles and pulse-charged CDI. We developed a theoretical model and

governing equations that would allow us to measure internal ion drift rates from externally

observable variables. We then designed, built, and tested a prototype CDI cell capable of operating

in either mode. We performed experimental work, as well as numerical simulations for

pulse-charged CDI over a range of realistic conditions, capturing a set of crucial timescales.

We find that practical restrictions required to avoid redox reactions in pulse-charged CDI

ensure that optimal powering profiles will always offer the greater benefit.

Providing safe drinking water access in the developing world will require a new generation

of technologies. These must be paired, however, with an improved understanding of how to

implement such projects, such that infrastructure proves resilient and impacts prove permanent.

I am optimistic that the work presented here may contribute, incrementally, to each

of these efforts.