manuvanegas Goto Github PK

I am a purpose-driven researcher who is eager to contribute to a rapid transition into a low carbon, resilient, and just society.

In my journey of finding ways to make a difference, I have used models and simulations to tackle problems like bioremediation, bioplastic production, water provision in a mega-city, agroforestry-based coffee growing systems that help the farmers reduce their losses to Coffee Leaf Rust, and, lately, emerging climate intervention technologies and the effects of their implementation in response to local climate change impacts. In all, I accrue 7+ years of experience modeling biological and socio-ecological systems. I have developed a unique skillset that blends micro/biology and ecology literacy with spatially-explicit agent-based modeling, parallel computing, container technologies, scripting languages, and quantitative analysis of large datasets.

I am an effective learner that enjoys solving puzzles, thinking in systems, and creating solutions. Currently, I am looking for opportunities where I can apply and grow my skills to help address the climate crisis.

As part of my recently completed doctoral research (dissertation defended in Summer, 2023, and degree to be conferred in December of this year), I developed SpatialRust, a spatially-explicit agent-based model. SpatialRust simulates the development of Coffee Leaf Rust in a coffee farm under conventional or agroforestry-based farm management approaches. I developed this model from scratch by distilling previous coffee plant growth and yield models, summarizing the existing Coffee Leaf Rust-related physiology, ecology and epidiemiology knowledge, and simulating thousands of these individuals in a stochastic and spatially-explicit environment.

Using this model, I tested different management patterns (varying aspects like the investment on different Coffee Rust control efforts and the spatial placement and management of shade trees) to find the ones that are more likely to maximize the farm's profitability. To explore this optimization problem, I implemented a Genetic Algorithm to efficiently explore the parameter space and find a farm management pattern that maximized the short- and medium-term profitability of the farm. Finally, I expanded the spatial extent of the simulations to cover more than one farm, which allowed to study the collective action problem related to the potential trade-off between short-term profits and beneficial long-term Coffee Leaf Rust management strategies.

The model code was archived under the following doi: https://doi.org/10.5281/zenodo.8237934 and is available to use under an MIT license. I am in the process of preparing a manuscript based on some of the contents of the dissertation, which I'll share here as soon as possible, along with the link to the dissertation itself, which has not been published yet!

As part of my work at CoMSES Net and its goal of supporting transparent, interoperable, and reusable scientific computation in the study of complex social and ecological systems, I have contributed with content and tools to advance the adoption of containerization technologies in the computational modelling community. We published a peer-reviewed article introducing containerization concepts to the socio-environmental modeling community and are developing a series of GitHub-based containerization tutorials (see the Containerizing an R Model and Containerizing an NetLogo Model classrooms).