Spatio-Temporal Analysis of CO2 Emissions

a Supper & Supper Use Case

Project Goal

Monitoring and reporting of CO2 emissions has a decisive impact on the negotiations of global climate change. However, the current system refers to non-standardized national reports based only on industry specific statistics. This results in non-verifiable estimates of national CO2 emissions. The project goal was therefore to identify alternative methods to make global CO2 emissions quantifiable and objectively verifiable.

The core dataset included measurements of surface CO2 concentrations obtained from Japan’s Greenhouse Gases Observing Satellite (GOSAT) comprising the majority of the landmass of Europe and Asia. As supporting datasets NASA satellite measurements of vegetation indices, population density, carbon monoxide and nitrogen dioxide were used.

The first challenge addresses the complexity of the data as high dimensionality in space and time quickly leads to bottlenecks in computational resources. This was solved with approximation approaches that analyzed the trade-off between predictive performance and computational demands.

The second task deals with providing a model framework for the CO2 concentrations, that allows for a space-time prediction of the sparse dataset and uses this correlation structure within the data. Spatial modelling captured the space-time dynamics.

The major challenge was to identify the CO2 sources and sinks as well as to find out how much CO2 has been emitted and how much variation in the surface CO2 concentrations relates to terrestrial vegetation or is caused by the population. The complete carbon cycle was integrated in the model framework including seasonally varying vegetation.

In the end, it was possible to quantify CO2 emissions caused by mankind.

An alternative to the existing approach of monitoring and reporting CO2 emissions was developed. It is capable of efficiently mapping the CO2 concentration process globally and the whole terrestrial CO2 cycle based on objective satellite measurements of CO2 emissions. A comparative study showed that the model results were largely consistent with the reported emissions (UNFCCC) within European countries, but differed widely for countries such as China and North Korea. This indicates an error (intentional or not) in the currently implemented reporting system.

← back to use case overview

Category

→ GEO AI
Spatial temporal Analysis

Technologies

Spatio-temporal statistics

Download

Social Sharing

Other Use Cases in this category