Needed input data to calculate N2O emissions
Categories of input data
We need several types of inputs. The on-field emissions depend on:
Further, to quantify emission intensities, we need agricultural data:
- fuel energy content per kg of crop
- area on which the crop is grown
- yield of the crop
- nitrogen inputs (fertilizer, manure, crop residues)
And finally, to determine the emission intensity for the biofuel consumed in the EU, we also need trade statistics for the crops. For now, only crop trade is included, a further improvement would be to also include biofuel trade.
Climate data
Dry or wet climate
We use IPCC data from JRC ESDAC to define climate zones. For the formulae, we need to know if the climate is dry or tropical.
Precipitation and evapotranspiration
To determine if leaching is occuring, IPCC uses precipitation and evapotranspiration to determine if the water remaining in the soil is more than the water retention capacities for months that are defined as a rainy season.
We use MODIS data for potential evapotranspiration as a proxy for pan evapotranspiration, and ECMWF ERA5 data for precipitation. As the first layer is an 8 day sum, we aggregate the precipitation also as 8 day sum.
For each period, we determine if it's a rainy season: "rainy season(s) can be taken as the period(s) when rainfall > 0.5 * Pan Evaporation" (IPCC 2019).
For all periods that is defined as rainy season, we sum the precipitation and subtract the potential evapotranspiration: remaining water = Σ(rain in rainyseason) - Σ (PE in same period).
Then leaching is occuring if this value is greater than the water retention capacity: leaching = remaining water > soil water holding capacity.
To align and rasterize the data, we use google earth engine.
Soil water retention is explained in soil.
Soil data
Organic soils
IPCC 2019 equation 11.1 defines N2O emissions from drained/managed organic soils. To determine the share of organic soils used for the cultivation of crops potentially used for biofuels, we use the FAO dataset Organic soils (histosols) and drained organic soils.
To our understanding, the layer "histoshare" contains the fraction of each pixel that is covered by organic soils, and "share_hist_crop_grass" contains the fraction of the organic soil used for crops (or grass). So to obtain the share of the pixel that is occupied by crops on organic soil, we multiply both layers. When multiplying with the area in each cell occupied by each crop, we obtain the area of each crop on organic soils in this pixel. This contains the assumption that in each pixel, the share of each crop on organic soils is the same.
Soil water retention capacity
To determine if leaching occurs, we need to know the soil water retention capacity, which we obtain from the Harmonized World Soil Database (HWSD). The data is also available in google earth engine as "AWC".
Agricutural data
General
The agricultural inputs are used to determine the area on which crops are grown and the nitrogen inputs to the fields. Crops analysed in this tool are all main crops used for biofuels:
| UNFAO_aggregated |
|---|
| Maize |
| Sugar beet |
| Wheat |
| Rapeseed |
| Soybeans |
| Sunflower seed |
| Oil palm fruit |
| Sugar cane |
| Maize, green |
| Maize, for silage |
Crop area
In the original tool, a publication from Monfreda et al. 2008 was used. It used an algorithm to spatially distribute the harvested area for each crop, based on national and sub-national statistics, land occupation maps and other assumptions.
To update the values, this tool uses CROPGRIDS. Currently, the preprint has been withdrawn, and publication in a different journal is in progress.
In CROPGRIDS, there are 3 categories for maize. As this is the only dataset containing "maize, for silage", we group it with the data for "maize, green".
Crop yields
In the original tool, a publication from Monfreda et al. 2008 was used. As there is no similar updated study to our knowledge, the spatial distribution information of this study is kept.
As CROPGRIDS harvested areas are no longer aligned with Monfreda et al. yields, the yields have to be extended to newly added harvested areas. To do so, the following approaach has been implemented:
- Country-wise crop yields from FAO are downloaded from FAOstat.
- Monfreda yield is extended by filling all areas with missing data by average country-wide yield from FAO
- Production based on (extrapolated) Monfreda yield multiplied with CROPGRIDS harvested area is calculated and then summed up for each country.
- This production is divided by the countrywise crop area from CROPGRIDS to obtain an average yield.
- Extrapolated Monfreda yield is scaled with the quotient of FAO countrywise yield divided by the above calculated average yield.
Finally, to avoid unrealistic values because of mismatches between crop area and yield from different sources, we limit the maximum new yield to 110% of the previous Monfreda yield. This is because in some countries with low production of a certain crop, the above algorithm obtained sometimes yield more than two times higher than the best Monfreda yield.
Mineral fertilizer inputs
Fertilizer inputs per country and crop are reported by FUBC.
This study is regularly updated, but the crop categories changed in the last version. We manually checked the evolution of fertilizer use per area for selected crops and could observe a decreasing tendency in the first reports and an increase for some crops in the last report. Cross-checking with FAOstat data which doesn't contain values per crop, we concluded that the change in fertilizer inputs doesn't justify an error-prone averaging or scaling, so we decided to use the original data from the FUBC 9 report (published in 2022, values from around 2018).
Warning
For some countries, fertilizer input for rapeseed is "hidden" in the category "Other oil Crops" (especially Australia, a major exporter of rapeseed to Europe). For those cases, we use the value of "Other oil Crops".
Warning
Around 10% of EU imports of palm oil come from countries where FUBC doesn't provide a fertilizer input value. To correctly take them into account, we use the same value as for neighboring countries (Mexico for south american countries, Indonesia for Papa New Guinea).
Manure inputs
In the old version of GNOC, manure inputs came from the EDGAR tool. To keep consistency of data sources, we decided to take values from FAOstat data. For each country, we calculate the ratio of manure applied to soils divided by mineral fertilizers applied to soils based on FAOstat data over the last 10 years.
This factor is then multiplied by the mineral fertilizer input per crop and country from FUBC to obtain the manure input per crop and country.
Crop residues
IPCC formulae use the fraction of crop residues removed from the field and the fraction of crop residues burnt on field. We take those values from the old GNOC tool as we didn't find a more reliable data source.
Fuel energy content per kg of crop
Lastly, to convert the N2O emissions into emissions per MJ of fuel,
| fertcropgr | LHV_MJ_per_kg_dry_biomass |
|---|---|
| barley | 17 |
| cassava | 16.15 |
| coconut | 32.07 |
| cotton | 22.64 |
| maize | 17.3 |
| oilpalm | 24 |
| ooilseeds | 25.9 |
| othcereal | 17.1 |
| othcereal | 16.9 |
| rapeseed | 26.976 |
| sorghum | 17.3 |
| soybean | 23 |
| sugarbeet | 16.3 |
| sugarcane | 19.6 |
| sunflower | 26.4 |
| wheat | 17 |