Regional greenhouse gas net emission intensities by land cover category in Finland

The methods related to the data published herein are described in detail in the associated publications (Holmberg et al. 2023, Junttila et al. 2023). This file describes the datasets and the data preparation steps. The aim of this data publication is to provide regional assessments of the role of land cover in greenhouse gas emissions in Finland. The results in the publications are reported for the large administrative divisions, the NUTS 3 regions of mainland Finland (Statistics Finland 2023a). While limited by the accuracy of the methods and source data involved, these data can also be used for more local assessments, e.g., at the scale of municipalities. The data represent a temporal snapshot of land cover. Except for the soil maps, rivers and lakes, all land cover data are from the period 2015-2020 and are based on registry data or remote sensing.

Data description

Data format. The data are distributed as GeoTiff raster files, which can be read using most GIS-software.

Units and definitions. The land cover net emission intensities are shared as raster data with a 250m-by-250m resolution in the ETRS-TM35FIN projected coordinate system. Negative values correspond to sinks (only sinks of C/CO₂ considered). The emission intensities are reported as total emission intensities in carbon dioxide equivalents (gCO₂-eq m^-2) based on the 100-year global warming potential as reported in the IPCC 5^th assessment report (Myhre et al. 2013, p. 73). All cells which do not include emissions from the corresponding land use are classified as NULLs or no data, which should be taken into account if combining raster layers. Where the source data report emission coefficients in the amount of the main element (e.g. C for CO₂ or N for N₂O) they have been converted to the amounts of the corresponding gas using the standard atomic weights of the relevant atoms (C: 12.011, O: 15.999, N: 14.007, H: 1.008) before conversion to carbon dioxide equivalents. See the related publication for the values of the emission coefficients used and further methodological details (Holmberg et al. 2023).

Data processing. Data processing for the production of the 250m-by-250m emission intensity raster maps was conducted using GRASS GIS 8.2 (GRASS Development Team, 2022).

Land cover emissions derived from vector data (rivers, lakes, agricultural land) were rasterized at a resolution of 1m² with the emission intensity as the raster cell value. For rivers, linear features representing rivers having a width of 2 to 5 meters were converted first to areal features by creating a buffer of 1.75 meters to represent an average width of 3.5 meters (see Rivers below). The buffer was created without caps so that the total length of the linear segments was not changed. The buffered river features were merged with the areal features removing the potentially overlapping parts.

For all source raster data, the data were available at a 16m-by-16m meter resolution. Emission intensities were aggregated to 250m-by-250m by first summing over the original raster cells intersecting with each aggregate cell while accounting for the proportion of each cell overlapping with the aggregated cell and then multiplying by the area of the original cell. The resulting raster values were divided by the total area of the aggregate cell to acquire average emission intensities. Hence, rasters including a lower proportion of the corresponding land use have lower emission intensities.

Thematic layers

Cropland. CO₂ emissions from cropland were estimated for mineral soils and organic soils separately using emission coefficients from the national greenhouse gas inventory report for 2023. Averaged emission coefficients for the years 2010–2020 for southern and northern Finland were used for mineral soils (Statistics Finland 2023b, Table 3_App_6j). For organic soils separate emission coefficients were used for annual and perennial crops (IPCC 2014, Table 2.1). Cropland and crop data were acquired from the Finnish Food Authority’s Land parcel register for year 2020. Soils were classified into mineral and organic soils by intersecting the field parcels with the soil body layer of the Finnish soil database (Lilja et al. 2006, Lilja et al. 2017).

Data files:

• Net missions from cropland on mineral soils: cropland_mineral_250m_250m_mean.tif
• Net missions from cropland on organic soils: cropland_organic_250m_250m_mean.tif

Forests. The net emissions from forests are estimated as the balance of carbon sequestration due to gross primary production of trees and understory vegetation and carbon loss due to harvested biomass, and emission from decomposition of harvest residues, litter, and soil organic matter. Forest productivity is modelled using the process-based forest growth model PREBAS (Minunno et al. 2016, 2019, Junttila et al. 2023, Mäkelä et al. 2023). The initial state for the forest model for the three main forestry species Scots pine, Norway spruce, and Silver birch is derived from the multi-source national forest inventory (MS-NFI; version 2015) and harvesting intensities are modelled on the basis of the Finnish national statistics (National Resources Institute Finland 2023). The PREBAS forest net emissions represent annual averages for the period 2017–2025.

CO₂ emissions from decomposition on mineral soils are estimated with the soil carbon model YASSO07 (Liski et al. 2005, Tuomi et al. 2009). On drained peatlands, in addition to CO₂ emissions due to peat and litter decomposition, the soil emissions include the CH₄ and N₂O emissions. The net emissions due to CO₂, CH₄ and N₂O from drained peatland (Ojanen et al. 2010, Ojanen and Minkkinen 2019, Minkkinen et al. 2020, Junttila et al. 2023) are calculated using emission coefficients for nutrient rich sites (herb-rich and blueberry type), and nutrient poor sites (lingonberry, dwarf-shrub, and lichen type).

Data files:

• Net missions from forest on mineral soils: forest_mineral_250m_250m_mean.tif
• Net missions from forest on organic soils: forest_organic_250m_250m_mean.tif

Lakes. Emissions of CO₂ and CH₄ were estimated for lakes using size dependent emission coefficients. The lakes were classified into five size classes with emission coefficients for CO₂ evasion (Kortelainen et al. 2006), CH₄ diffusion (Juutinen et al. 2009) and ebullition (Bastviken et al. 2004) as well as the CH₄ emissions due to the macrophytes Phragmites australis and Equisetum fluviatile (Juutinen et al. 2003, Bergström et al. 2007, 2011). The lake date was from the Shoreline10 data by the Finnish Environment Institute.

Data files:

• Net emissions from lakes: lakes_250m_250m_mean.tif

Rivers. CO₂ emissions from rivers were estimated using emission coefficients based on the width of the stream. The width dependent emission coefficients were derived from stream order specific emission coefficients of Swedish rivers (Humborg et al. 2010) by classifying the rivers into width groups and with the emission coefficients chosen based on the stream order specific coefficient of corresponding average width. The river emissions were calculated from the Shoreline10 data by the Finnish Environment Institute which represents rivers wider than 5 m as areal features, and rivers < 5 m wide as linear features. For rivers < 5m wide, an average width of 3.5 m was assumed.

Data files:

• Net emssions from rivers: rivers_250m_250m_mean.tif

Undrained mires. Total net emissions were estimated for undrained mires in Finland using average emission coefficients for CH₄ (Minkkinen and Ojanen 2013), CO₂ (Sallantaus 1994 , Turunen et al. 2002), and N₂O (Minkkinen et al. 2020). The emission coefficients represent the long term accumulation of carbon as well as the emission of CH₄ and from N₂O peatland. Peatland sites were extracted from the multi-source national forest inventory (MS-NFI; version 2019; see also Mäkisara et al. 2022) and undrained mires were delineated using data provided by the Natural Resources Institute Finland. The undrained mires were classified into four classes using the MS-NFI data: 1) productive forested mires, 2) sedge fens, 3) other open and sparsely treed fens and 4) ombrotrophic bogs, which mainly differ in their emission coefficients for methane (Minkkinen and Ojanen 2013).

Data files:

• Net missions from undrained mires: undrained_mires_250m_250m_mean.tif