UC Berkeley

Today, 3 billion people, or 41% of the global population, burn wood, charcoal, dung, crop residues, coal, or other solid fuels to cook their food. This practice has resulted in significant loss of human life and environmental damage. In fact, smoke from traditional cooking fires is one of, if not the, most important environmental risk factor to human health today; recent estimates attribute some 4 million annual premature deaths to emissions from cooking. While the total proportion of the global population cooking on solid fuels decreased from 62% to 41% in the three decades from 1980 to 2010, a growing global population means the total number of people affected by traditional cooking has remained essentially unchanged. Many experts estimate that this state of affairs—three billion people cooking on solid fuels—will remain unchanged in our lifetimes. In addition to cooking’s toll on human health, traditional cooking has a substantial environmental impact. From forest denudation to emissions of climate-warming gases and aerosols, traditional cooking has impacts far beyond the indoor environment.

To ameliorate this crisis, many stakeholders have turned to “clean cookstoves.” The cookstoves, which typically burn locally-available solid fuels, aim to deliver the same quality and quantity of cooking as traditional cookstoves but with less fuel and fewer emissions. The net impact of one of these cookstoves will be some function of of three important criteria: how well the cookstove performs (emissions), how much the cookstove is utilized (adoption), and how effective the new cookstove is in curtailing emissions of the traditional baseline cookstove (replacement).

Considering the enormity of this crisis, relatively little research has been done to validate the impact of cookstoves. In the past, first-generation cookstove monitoring and evaluation relied on surveys and direct observation as the primary tools to evaluate adoption. These validation techniques were insufficient for several reasons. Firstly, surveys suffer from recall bias; participants cannot recall their exact behavior even if they desire to do so. Secondly, respondents may misrepresent their actual behavior to enumerators in an effort to portray behavior that is more “desirable.” Finally, observed behavior suffers from the “Hawthorne Effect” in which subjects under observation modify their behavior as a result of the salience of observation. In addition to the intrinsic limitations of these methods, surveys and observation cannot be used to cost-effectively monitor large populations with high temporal fidelity over long time periods.

In this dissertation, I discuss several case studies of second-generation monitoring and evaluation techniques. The work begins with a life cycle assessment of a particular cookstove, the Berkeley-Darfur Stove. This life cycle assessment bounds the sources of environmental impacts of the Berkeley-Darfur Stove over five life cycle phases: materials, manufacturing, transportation, use, and disposal. The use phase accounts for more than 99% of total lifetime CO2-equivalent emissions, but the Berkeley-Darfur Stove still emits less CO2-equivalent over its lifetime than a traditional fire. A significant proportion of Berkeley-Darfur Stove’s use phase CO2-equivalent emissions are due to products of incomplete combustion such as black carbon particulate matter; many of the same products of incomplete combustion that warm the climate are also responsible for damage to human health. This work concludes that future work on impact must focus on the use phase; almost any manufacturing material, method, or location would be justified from a lifetime CO2-equivalence perspective so long as the new cookstove has reduced per-meal emissions relative to the baseline cookstove and that the new cookstove replaces the baseline cookstove to some extent.

The next part of this work focuses on a deployment of stove use monitors (SUMs) in Darfur. In this study, we used sensors and surveys to measure objective versus self-reported adoption of freely-distributed cookstoves in Darfur, Sudan. Our data insights demonstrate how to effectively measure and promote adoption, especially in a humanitarian crisis. With sensors, we measured a 71% initial adoption rate compared to a 95% rate reported during surveys. No line of survey questioning, whether direct or indirect, predicted sensor-measured usage. For participants who rarely or never used their cookstoves after initial dissemination (“non-users”), we found significant increases in adoption after a simple followup survey (p = 0.001). The followup converted 83% of prior “non-users” to “users” with average daily adoption of 1.7 cooking hours over 2.2 meals. This increased adoption, which we posit resulted from cookstove familiarization and social conformity, was sustained for a 2-week observation period post intervention.

Unlike the Berkeley-Darfur Stove, which is a simple naturally-drafted cookstove, a new generation of forced-air cookstoves are coming onto the market. The most common technique for forcing air is with an electric fan powered by a thermo-electric generator (TEG). These TEGs produce electricity by converting the thermal power through a temperature gradient into electricity via the Seebeck effect. Many new TEG cookstoves promise higher rates of adoption by cooking faster, with less smoke, and, importantly, with outboard electricity provided via a USB port. The electrical power provided through these USB ports is small, but enough to charge a cell phone or power LED lights—a major boon to people living off grid. However, there has been some concern that these USB-enabled cookstoves will incentivize customers to burn fire for the sole purpose of charging phones without any traditional fire replacement. In the third part of this dissertation, I discuss a study of Advanced Stove Use Monitors (ASUMs) which were designed to monitor this phenomenon. The study, which took place in Odisha, India, involved 72 participants. We found that access to USB power dramatically improved adoption of the cookstove as a cooking tool (2.8X increased cooking adoption vs. an identical cookstove without USB power) at the expense of roughly 17% of total usage time being for charging only. This study goes on to conclude that this relatively small proportion of total use time used for charging-only is unlikely to negate the emissions exposure benefits of a cleaner cookstove.

Finally, we pull back from looking at per-user adoption and focus on macro-level indicators of solid fuel combustion, namely atmospheric aerosol of black carbon. Globally, black carbon accounts for about 25% of net radiative forcing, and black carbon from residential solid fuel cooking represents about 30% of the total atmospheric black carbon burden. Un- certainty in black carbon’s total contribution to radiative forcing is substantially higher than other species’ contributions because black carbon has a short atmospheric lifetime, is highly spatially heterogeneous, and its influence is highly dependent on its altitude and relationships to the albedo of its surroundings. We believe that a low-cost balloon-borne black carbon monitoring platform would materially increase scientific understanding of black carbon’s role in the climate change story while atmospheric profiles might also serve as a proxy for where and how much solid fuel is being combusted. This would allow researchers to use in situ profiles of black carbon as a validation for less accurate remote sensing instruments (e.g. satellites) while also serving as a basis for longitudinal studies of decreasing or increasing ground-based black carbon sources. In our work, we developed and deployed a black carbon sensor that can be carried on a balloon and can measure black carbon at a resolution that is comparable to commercial instruments. In this dissertation, I discuss the development, testing, validation, and test flights of the balloon-borne black carbon monitoring system.

In summary, this work moves towards next-generation monitoring and evaluation of the global crisis of cooking. Results point towards the need for in situ monitoring of cookstoves as well as promising methods for remote sensing of regional-sale biomass combustion. I present insights about how thermoelectric generators and USB ports can dramatically in- crease cookstove adoption, while highlighting that, in one study, survey-based measurement of adoption is of almost no value when quantifying adoption of cookstoves. This dissertation points to a future “Third Generation” of monitoring and evaluation that must focus heavily on networks of on-the-ground sensors, remote sensors, and implementation practices that improve adoption by closing the loop between implementers’ choices and measured user behaviors.