Summer Research | ENVIRONMENT and SOCIETY: DATA SCIENCES for the 21st CENTURY (DS421)

During the first summer of the program, trainees participate in Research Experience for Trainees (RET). Working with faculty, postdocs or senior graduate students in a complementary disciplines, trainees conduct a discrete, independent project and assist the mentor’s research. Trainees work in their new research group for 4 weeks (160 hours) allocated over the summer.

Here are some summaries of the experiences Cohort 1 and 2.

Jacob Bukoski Environmental Science, Policy and Management	Jacob worked with Dan Hammer — an award winning environmental data scientist — to rapidly assess timely research questions using remote sensing applications. More specifically, they used the Google Earth Engine platform to investigate real-time expansion of fossil fuel and mining interests on United States public land, and rapid military expansion on islands in the South China Sea. The work exhibits the value of real-time remote sensing to monitor and raise public awareness around environmentally or socially contentious events.
Nathaniel Decker City and Regional Planning	Nat worked with Chris Jones in ERG on a project that estimated the order-of-magnitude GHG impact of the spatial distribution of households by incomes in new residential construction.
Gabriel Englander Agriculture and Resource Economics	Gabe I helped Professor Laurel Larsen run simulations to explore a core question in ecology: Why do so many species coexist when it seems that many occupy the same ecological niche? Using stochastic tournaments, we find that spatial heterogeneity in competitive abilities is the key to explaining the high levels of biodiversity we observe in the real world.
Frank Errickson Energy Resources Group	Frank worked with Marc Fleurbaey from Princeton University on modeling health benefits of COs mitigation.
Scott Kaplan Agriculture and Resource Economics	Scott worked with Professor Joan Walker in the Department of Civil and Environmental Engineering. Joan and Scott were interested in understanding how the introduction of autonomous vehicles (AVs) would affect societal preferences for private vehicle ownership versus ride-sharing. Specifically, the research attempted to address whether the private-ownership model or the sharing model is likely to prevail in the case of AVs and to determine the factors that may affect movement towards one model versus the other.
Eli Lazarus Energy and Resources Group	Eli worked on a project initiated by theoretical ecologist John Harte, and supervised by him. This project is experimental theoretical economics using the techniques and insights from theoretical ecology. John Harte and PhD student Ann Kinzig explained the evolution of a dynamic of combined mutualism and competition between plants and decomposing microorganisms which describes the way the essential nutrient of nitrogen is shared. I worked on developing a macro economic version that models the same dynamic with macro economic agents in parallel to the plants and microorganisms dynamic. In collaboration with 2 other researchers, I built on a version where labor and capital share and also compete for resources, where the optimal allocation is determined by capital.
Niklas Lollo Energy and Resources Group	I am a second year MS/PhD student in the Energy and Resources Group. I plan to examine social and environmental indicators as tools for global governance, weaving together sociology of quantification, legal studies and data science. I have particular interest in understanding the tension between high-level summaries and low-level, action-enabling statistics.
Evangeline McGlynn Geography	Eve supported a project an ecology of borders, working with Jamon Van Den Hoek at OSU. Starting with a remote sensing driven analysis of ecotones, she intended to catalogue and compare ecosystem edges on the African continent and the political boundaries that either bisect or follow them. Moving forward from there, we hope to create a typology of borders, interpreting the border as a biophysical or biopolitical phenomenon rather than just socio-political infrastructure.
Rosanna Neuhausler Geography	Over the summer I worked with Maria Bruna and Martin Robinson at the University of Oxford within the Mathematical Institute and Department of Computer Science. Together we developed a stochastic spatio-temporal model of coastal benthic coverages as a thought experiment asking how to model meta-stability between ecological regime shifts. Specifically, we looked into the dynamics between coral, macroalgae, and algae turf based on herbivorous grazers, water quality, and coral demographics. We represent these actors through object-oriented programming using a python front end and C++ backend simulation software, Aboria. Grazers, included as individual fish (particles), evolve according to a kinetic model and interact with neighboring benthic assemblages, represented as nodes. Water quality and coral demographics are input parameters that can vary over time, allowing our model to be run for temporally changing scenarios and to be adjusted for different reefs. I presented our product at the European Coral Reef Symposium the following winter and am currently making a more documented version of the code available on GitHub.
Katherine Siegel Environmental Science, Policy and Management	Katherine worked with Dr. Lauren Ponisio (formerly a post doc at the Berkeley Institute for Data Science and now an assistant professor at UC Riverside) on studying the relationships between farming practices, crop yields, and food security in smallholder farms in Sub-Saharan Africa. This project was part of a collaboration with Conservation International’s Vital Signs initiative, using data collected by CI on environmental variables, farming practices and crop yields, and household characteristics across Ghana, Rwanda, Tanzania, and Uganda.
Carrie Tribble Integrative Biology	Carrie used a novel combined metric of greenness and thermal signatures developed by Dr. Iryna Dronova and examine the relationship between this metric and race of residents in Northern California urban centers such as San Francisco and Sacramento.

Laura Alexander Integrative Biology	Laura worked with the Getz lab using niche modeling to predict the range of Ebola using a dataset including plausible but unconfirmed outbreaks in humans and animals.
Jenna Baughman Integrative Biology	The Florida Everglades have been described as having regular or periodic landscape patterning, with characteristic bands of sawgrass-dominated ridges and slow-flowing sloughs. Yet, the processes that develop and maintain the Everglades are not well understood. Quantifying the spatial characteristics of these landscape patterns can provide insight into the processes as well as provide metrics for model validation. Using Matlab, I used spatial spectral analyses to compare landscape patterns and frequencies represented in simulated ridges and sloughs from the RASCAL Everglades model to observations from arial imagery. Worked with Laurel Larsen in Geography.
Ian Bolliger Energy and Resources Group	Solar Radiation Management (SRM) has been proposed as a potential geoengineering strategy to offset rising global temperatures. This approach would involve the injection of sulfate aerosols into the stratosphere, which would reflect sunlight and cool the Earth. While the resulting effect on global temperature is relatively understood, secondary local impacts on climate and weather patterns are largely unknown, though several regional impacts have been hypothesized. Regional climate changes could have profound impacts on agriculture, a spatially heterogeneous sector that is strongly tied to precipitation, temperature, and sunlight. In the study I participated in for the RET, we attempted to estimate the impact of proposed SRM scenarios on the yields of several staple crops. To do this, we first estimated the impact of a set of climate variables on yields, using large volcanic eruptions, which also spew sulfur into the stratosphere, as natural analogues. Next, we used the results of climate models in which SRM was and was not applied. This allowed us to identify the location-specific changes in the relevant climate variables that were due to SRM implementation. Applying the relationships identified in the first step to the model results from the second, we were able to estimate a yield impact of SRM. My particular work on this project involved the development of an analysis plan for the climate model data, model selection and data acquisition, and initial preprocessing of the model output.
Hal Gordon Agriculture and Resource Economics	Hal worked with Felipe Castro in the Energy and Resources Group on consumer electricity demand.
Andrew Hultgren Agriculture and Resource Economics	Andy worked with Duncan Calloway from the Energy Resources Group on the effect of the load of charging many electric vehicles on the energy grid.
Yang Ju Landscape Architecture and Environmental Planning	Yang worked with a graduate student in urban planning.
Matthew Kling Integrative Biology	I worked in Sol Hsiang’s lab at the Goldman School of Public Policy on a project collaborating with his PhD student Jon Kadish. The project was aimed at estimating the impacts of climate changes on electricity use in China. We combined historic data on localized electricity use, climate variation, changing appliance ownership, and population in a set of regression models, leveraging temperature variation across years and across Chinese provinces to estimate the energy impacts of changes in the frequency of hot and cold days. We were able to quantify how temperature anomalies have affected air conditioner purchases and electricity use in recent decades, helping to inform predictions about the impacts of future climate change on energy use.
Kate Pennington Agriculture and Resource Economics	Kate spent the summer building a data set that codes changes in vehicle emissions regulation in India from 1987-2016 at the district level. She plans to use the data to evaluate the impact of emissions policy on particulate matter concentrations, estimated emissions, local precipitation, crop yields, and health. Kate worked with John Chiang in Geography.
Dana Seidel Environmental Science, Policy and Management	For my RET project, I partnered with Dr. Gabriel Lenz, a professor in the Political Science department at UC Berkeley. The objective of my project was to explore the relationship between energy prices and the incumbent party’s vote share in presidential elections from 1970-2012. Energy price data was gathered for each state and year from the US Energy Information Administration and data on US voter participation by county and state were compiled by Dr. Lenz. Although I had previous experience analyzing large data sets of the course of my research, being a natural scientist in training, this was a new opportunity for me to work with data regarding human behavior and the first time I had assessed something so historical in nature. Over the course of this project, I gained a lot of experience cleaning and organizing multidimensional data from different sources in R. Given the complexity of the available data, I became very practiced in tools for exploratory analyses and for pattern visualization in the data as we explored early hypotheses. Notably, I learned a lot about how political scientists approach regression models and how this differs from the approaches I was more familiar with from the life sciences. The project offered me a unique opportunity to gain experience with data and methods in political science and to sharpen my general data analysis, visualization, and communication skills.
Sara Stoudt Statistics	Understanding the dynamics of water quality in the San Francisco Estuary: a Generalized Additive Model approach Long-term time series of water quality measurements spanning multiple locations in the San Francisco Estuary are available. The challenge is to leverage this data to learn how nutrient concentrations vary across space and time in the Delta, what the major drivers of this variability are, and if any mechanistic insights can be gained by exploring the relationships between various variables collected. Working with the San Francisco Estuary Institute (SFEI), I built and visualized several different models to provide insight into the complicated system of the Delta. Generalized Additive Models, if built in a nested way, can be easily visualized to see how each variable contributes to the overall variability in the quantity of interest. By creating a web application to easily explore and visualize the results from different models, scientists at SFEI can better identify potential drivers of change in the Delta and form hypotheses that can be tested with further analyses.
Carmen Tubbesing Environmental Science, Policy and Management	The epic drought that hit California between 2012 and 2016 has had devastating effects on our state’s forests, killing over 100 million trees. The state is scrambling to find ways to remove the dead trees left in the drought’s wake in order to reduce fire and safety hazards. However, only a small fraction of the dead trees could be processed by California’s few lumber mills and biomass energy facilities. Converting the dead tree biomass to energy is particularly difficult given the high cost of transporting dead trees to far-off biomass facilities and low energy prices in return. One possible solution would be to gasify and burn the biomass on-site with small, portable biomass gasifier generators, such as those created by All Power Labs here in Berkeley. To assess the feasibility of such a solution, Carmen worked with Jose Lara in the Energy and Resources Group and other researchers to model the costs and revenue that would result from wide deployment of biomass gasifiers. Carmen’s contribution to the project included calculating spatially-explicit estimates of the amounts of biomass that have resulted from drought-induced tree mortality. She used a combination of aerial assessments of numbers of dead trees in polygons throughout the state and a rasterized model of live biomass and forest characteristics at fine resolution. Carmen’s estimates, along with information on distance to roads and substations, topography, and other factors, allowed for a detailed analysis of where gasifiers could be economically deployed.
Valeri Vasquez Energy and Resources Group	I contributed to the work of the Climate Impact Lab, an interdisciplinary team of data scientists, climate experts, and economists, by spending my RET in the Global Policy Laboratory at the Goldman School of Public Policy. The goal of the Climate Impact Lab’s broader project is to conduct a meta-analysis of the existing climate impacts literature in order to develop an empirical basis for the Social Cost of Carbon (SCC). My particular focus was on estimating the relationship between climate change – namely, temperature and precipitation shifts – and human conflict.