Portfolio

Oikobit is led by Gabriel Arellano. This portfolio brings together work delivered through Oikobit and relevant earlier collaborations led or co-developed by Gabriel.

fgeo2: forest analysis code easier to use and maintain

Context: ongoing consulting engagement between the Smithsonian Tropical Research Institute and Oikobit; ForestGEO analytical workflows.

Problem: forest plot researchers work with a large body of R code built over many years. Existing functions vary in relevance, clarity and maintainability. Network of dependencies is too vast and complex (see image) and makes many bugs virtually impossible to trace and fix by the average R user. New users need dependable ways to understand and reuse the methods.

Contribution: evaluate legacy methods and their dependencies, remove and simplify code where justified, develop focused functions, and provide tests and tutorials alongside the scientific reasoning. See github.com/gabrielareto/fgeo2.

ATFS: making forest datasets comparable across networks

Context: Alliance for Tropical Forest Science; Gabriel's networking role as co-PI, senior author, and coordinator of the data federation working group.

Problem: research networks use different data formats and definitions, and raw datasets may remain under local governance. Cross-network science needs agreed meanings and reproducible transformations.

Contribution: coordinate and contribute to horizontal data federation, quick harmonization of different datasets, and their scientific use. The associated DataHarmonization application is a collaboration with different credited developers. See alliancetropicalforestscience.net/working-groups.html and github.com/Alliance-for-Tropical-Forest-Science/DataHarmonization.

Madidi: tracing data errors from forest plots to specimens

Context: The Madidi Project linked forest-plot records in Bolivia to botanical specimens and taxonomic determinations. Reliable ecological analyses depended on those links producing consistent, traceable species information.

Problem: Missing determinations, inconsistent representative-specimen links and lost information about who established and surveyed plots could affect the resulting dataset. Some problems belonged in the source database; others could be handled in analysis code or required review by the people who knew the specimens and plots.

Contribution: Gabriel developed a scripted error report that grouped problems by type and identified the affected trees and specimens. The report gave collaborators specific records to review and indicated where corrections should be made, while keeping cases that could not be resolved automatically visible rather than silently assigning them a species name.

Local forest biomass for heating

Context: design of a system to supply municipal heating facilities with locally sourced forest wood in a region in northern Spain. Consultancy provided for Foresna Zurgaia; project promoted by TEDER (Tierra Estella - Desarrollo Rural).

Problem: Years of technical, economic and legal studies had produced a large but fragmented evidence base. Biomass supply and future demand remained uncertain, while the proposed arrangement had to work for net contributors and net buyers in a fair way. Without careful design, the benefits and costs could fall unevenly between them.

Contribution: Oikobit reviewed and synthesized more than 3700 pages of earlier documentation and commissions on forest supply, heating demand, logistics and organizational options. Then, Oikobit developed a proposed operating and incentive framework for municipal participation. The framework connected wood sales, processing services and chip purchases, and made economic viability and the distribution of benefits explicit matters for decision.

Forest crowns in three dimensions

Context: Forest inventories record tree locations, diameters and species, while aerial images show the canopy from above. Models of species coexistence need estimates of how trees compete with each other in three dimensions. Analytical consultancy work for the University of Washington.

Problem: Neither standard field inventories or aerial images provide estimates of how tree crowns are distributed, relative to each other, in three dimensions. In the absence of costly LiDAR data, allometrical approaches are needed.

Contribution: Gabriel developed an allometry-based model of tree crowns and used it in a three-dimensional analysis of the Luquillo forest in Puerto Rico. The resulting paper in Ecology showed that vertical separation between species was common, while three-dimensional aggregation was uncommon.

Product development: The model serves as a basis for a more complete machinery to produce a best-guess view of how the canopy is seen from above, using ground inventory data. This opens a route to reduce noise in the comparison between plot data and remote sensing observations. Validation against observed crowns and propagation of uncertainty is still ongoing work.

Reviews and analyses for MITECO plantations

Context: The MITECO program in Spain incentivizes tree plantations as part of long-term commitments to mitigate climate change. Different companies and investors act within this framework.

Problem: MITECO parameters are outdated and lagging behind European regulation. They are not ecologically realistic in many cases. On the other hand, they remain the legal basis for current project valuation. It is important for investors to design plantations to balance current incentives with more realistic calculations of biomass growth into the future. Fat-tail risks are particularly important in Spain, due to summer droughts and wildfire risks.

Contribution: Oikobit works actively on desktop reviews for prospective investors, and analytical products for realistic projections of biomass growth that take local conditions into consideration. See this evaluation as an example. Other work has been done under more strict non-disclosure agreements.

Carbon finance and forest conservation in Maniema

Context: Green House Capital was developing a forest conservation and sustainable development project in Maniema, DR Congo. Its proposed activities ranged from reducing deforestation to introducing efficient cookstoves, water filters and biochar. The team needed to understand which activities might qualify for carbon finance and what their potential climate benefits could be.

Problem: Credible carbon estimates require a clear project boundary, gather evidence about local land use and fuelwood consumption, and develop a realistic range of scenarios with limited information.

Contribution: In collaboration with Alicia Ledo (independent consultant), Oikobit developed reviewed possibilities in the context of carbon markets, and helped develop a roadmap for project design. We assembled available data to estimate existing carbon stocks and emissions, then compared a business-as-usual scenario with the proposed interventions while making assumptions and uncertainty explicit. The work helped identify promising activities and the mapping, household surveys and methodological checks needed before a formal carbon project could be developed.

ForestGEO: protocol for tree mortality and damage

Context: collaborative field implementation within ForestGEO and Next Generation Ecosystem Experiment - Tropics.

Problem: lack of understanding about patterns and processes in tree mortality in the tropics is a major source of uncertainty in Global Earth models. Additional field work, designed specifically for smooth incorporation into large models like FATES, is needed. Tree deaths and damage need to be observed and recorded in consistent ways across sites and field teams for subsequent analysis.

Contribution: Gabriel developed a repeatable field protocol and supports its use in research teams, linking observable categories to later data analysis. The work includes the full pipeline from field forms to usable datasets, including un-doing the stratification in the sample and going from quick field codes to full unequivocal fields in the datasets. See https://onlinelibrary.wiley.com/doi/10.1111/jvs.12981. Gabriel continues providing training for new field crews as an external consultant through Oikobit.

info@oikobit.com

Oikobit LLC. 2105 Vista Oeste St NW, Suite E - 1154. Albuquerque, New Mexico 87120, U.S.A.

Page updated

Google Sites

Report abuse