CO2 concentration monitoring: Defining sinks and sources using methods of image processing

Vortrag

Teil der Sitzung

[FS-446] Trade-offs and synergies between climate mitigation, biodiversity protection, food security and further sustainability goals (1/2)

Sitzungstermin

Freitag (22. September 2023), 09:00–10:30

Sitzungsraum

HZ 8

Autor*innen

Yana Savytska (Universität Tübingen)

Kira Rehfeld (Universität Tübingen)

Kurzbeschreibung

Afforestation and reforestation could be ways to reduce the atmospheric CO2 concentration. To assess effectiveness, and potentially enhance it, we identify areas with higher and lower CO2 fixation rates, and relate them to local vegetation diversity.

Schlagwörter

CO2 concentrations, afforestation, image processing, graphical filters.

Abstract

Biodiversity loss and planetary heating are currently accelerating due to human actions. Afforestation and reforestation are often discussed options to reduce CO2 concentrations (CDC). We aim to test co-benefits of protecting and promoting biodiversity, and enhancing climate mitigation through CO2 reduction. We identify regions, where high vegetation diversity exists, and carbon dioxide is effectively fixed.

In ecosystem CO2 is fixed with a biomass. Amount and properties of this biomass are different for different species. One single species ecosystem potentially fixes less CO2 than ecosystems with different species.

In our research we consider an ecosystem as a separate cell with a set of features. Taking into account the CO2 balance in cells with equal CO2 emission, the resulting concentration will be lower in cells with a better CO2 fixation. These we need to identify. To this end, we apply image filtering methods to CO2 datasets.

The changes of CDC levels we define as border of cells locations. This problem can be solved as an image recognition task. The main task is therefore to define areas with higher (or lower) CO2 fixation than in the neighboring areas.

For this purpose, Laplacians (high-frequency) filters are used. This filter has a transmission ratio equal to 0 (detect the difference) and “highlights” boundaries. From a logical point of view, the filtering operation can be presented as weighted difference between elements located around the “center” of the filter and the center element.

This filtering operation allows to find the position of CO2 sources and sinks. The source will be detected if the filter’s process result is greater than zero, and the sink will be detected if the result is less than zero.

The data obtained as a result of processing by the Laplacian filters are deviations of the CDC at the considered point in space from the average value of the CDC at neighboring points. The proposed algorithm can be used both for the localization of CO2 sources and sinks.

In the next step we aim to identify the vegetation composition and diversity in the sink regions in order to answer our main research question: “Can forests with higher biodiversity be more effective carbon sinks?”