Onderwerp scriptie

Exploring data quality to compute greenhouse gas emissions for major feed crops - Ellen Huls

As the demand for animal-source food is expected to double by 2050, the need to tackle climate change by reducing GHG emissions from livestock is becoming urgent. Feed production accounts for about 45% of sector emissions and, therefore, a sound assessment of the GHG emissions during the production of feed ingredients is of major importance to mitigate climate change. Our objectives were: to collect existing data and investigate the quality of available data to assess GHG emissions from feed production.

The global livestock sector is responsible for about 14.5% of the emissions of greenhouse gases (GHGs), mainly via carbon dioxide (CO2), methane (CH4), and nitrous oxide (N2O). As the demand for animal-source food is expected to double by 2050, the need to tackle climate change by reducing GHG emissions from livestock is becoming urgent. To identify mitigation options, quantification of GHG emissions along livestock production chains is required. Feed production accounts for about 45% of sector emissions and, therefore, a sound assessment of the GHG emissions during the production of feed ingredients is of major importance to mitigate climate change. Variability in data, however, can considerably affect this assessment and, therefore, a sound assessment requires a general accepted method and high quality data. Our objectives, therefore, were: to collect existing data and investigate the quality of available data to assess GHG emissions from feed production.

Five major feed crops were selected based on the highest production of feed globally: maize, soybeans, wheat, barley, and cassava. The crop system considered is includes the production stage of the feed ingredient, from the cradle to the arable farm, including the production of inputs, cultivation, and storage. Within this system we explored the relevant CO2, CH4, and N2O emissions and investigated how to calculate a carbon footprint of a feed ingredient to determine the essential data to include in our database. Data collection was based on literature, frequently used databases, and new databases recommended by experts. Data quality was assessed based on the Pedigree matrix of Ecoinvent, including criteria as: reliability, completeness, and temporal, geographical, and further technical representativeness.

Our database includes all data required to calculate the carbon footprint of a feed ingredient, given the system boundary of the study. We found data in 58 literature sources and 8 frequently used databases, in which data about production and application of inputs are more readily available than data about management practices and storage. Transparency in the uncertainty of data is necessary to reduce the uncertainty of data. By conducting a data quality assessment, the lack of representative data and data gaps are identified. The pedigree matrix is a simple and concrete method to assess data quality. Descriptions of the assessment criteria, however, appeared not evident enough to apply this method on every specific study. Using decision trees, which we defined ourselves and correspond with the data quality goal of the study, provides a concrete structure for assessing data quality.

Student: ELJM Huls

Supervisor: prof dr ir I de Boer

36 Ects