GSOC17 project

The final version of the Global Soil Organic Carbon Map (GSOCmap) is available on the Food and Agriculture Organization of the United Nations (FAO) website.

The map is the result of the work of researchers from different countries on the GSOC17 project of the Global Soil Partnership.

Each participating country prepared the map independently, based on the general principles detailed in the guidelines.

These are summarized here:

• Reserves for mineral and organic soil layers are calculated separately (with the possibility of dividing the organic layer into litter and peat layer). In the final map, these three layers were summed up;
• The input data for the calculations were values:
  1. organic carbon content in the layer up to 30 cm, % (for the section or contour of the soil map);
  2. volumetric mass in natural composition, g/cm³ (for the section or contour of the soil map);
  3. the degree of stoniness, % or gradation (for the cut or contour of the soil map);
• The result of the national map calculation was a grid with 1×1 km pitch (or, more precisely, ½th of an angle second or ¹/₁₂₀ degrees). The grid was calculated using conventional interpolation, digital soil mapping or a combination of these methods depending on the set of initial data - archive soil maps, soil profile descriptions, agrochemical surveys.
• In the same format, a validation map is provided, i.e. estimates of the error of calculation of the main map;
• Cross-linking of national maps into a single global map was carried out by the FAO Global Soil Partnership Secretariat together with ISRIC using common masks of water bodies, administrative boundaries, etc., due to which there may be slight differences in the total data for individual countries.

In domestic soil science, the assessment of organic carbon stocks in soils and their cartographic representation was carried out several times (D.S. Orlov et al., 1996; V.A. Rozhkov et al., 1997; Stolbovoi V., 2002; A.A. Titlyanova et al., 1998; D.G. Shchepashchenko et al., 2013; etc.). However, for such a country as Russia, with its vast territory and a huge variety of natural conditions, the construction of a carbon stock map in accordance with the declared GSOC17 program principles is possible only through coordinated joint efforts of the teams of many scientific and industrial organizations.

The Soil Data Center of the Lomonosov Moscow State University acted as the coordinator of the map preparation for the territory of Russia.

The Soil Data Center of MSU is the initiator of exchange of multi-scale soil data within the framework of distributed network of soil data processing centers, which is currently being created in Russia. All works are carried out on the basis of the Information System "Soil and Geographic Database of Russia" (IS SGDR), the capabilities of which correspond to the ideology of the FAO Global Soil Partnership Program.

IS SGDR allows to operate with a variety of data accumulated in different institutions: different-scale cartographic information, attribute data of geographically referenced objects, both point data - by soil profiles and area data - that characterize certain areas or territories.

IS SGDR also includes a vectorized map of soils, soil and ecological zoning, a map of forests and land use, and a map of landscapes, all at a scale of 1: 2.5 M (S. Shoba et al, 2011; Golozubov et al., 2015; Urusevskaya et al., 2015).

The main principles of drawing a map of organic carbon stocks in the 30 cm soil layer of Russia are described below.

In accordance with the concept of a distributed network of soil data processing and storage centers, the calculation of the map of organic reserves in the 30-cm soil layer is performed in the form of synthesis of several types of initial data:

a) maps of the entire territory of the Russian Federation based on the Soil map of the RSFSR (edited by V.M. Fridland, 1988), the Soil Geographical Zoning Map (edited by G.V. Dobrovolsky and I.S. Urusevskaya, 2013), landscapes, forests and land use (all on a scale of 1: 2.5M) in combination with a sparse irregular grid consisting of about 2000 soil profiles.

Calculations used: more than 25,000 polygons of the soil map, the legend includes more than 300 allotments. Along with soils and soil complexes, the legend also includes non-soil formations. C stock values for these compartments are assumed to be small fixed values to be able to recognize them and distinguish them from "no data" (see table below).

The principles of calculation of volume mass and carbon stocks in soils (for the first soils of the soil map legend) with and without the use of various pedotransfer functions, averaged values and based on expert estimates are described in the Appendix.

(b) Maps of selected agricultural areas based on large-scale soil maps and dense grid data of regular soil agrochemical surveys, obtained in agreement with the Ministry of Agriculture of the Russian Federation. Two soil organic C reserves maps were created for a limited area of about 15 million hectares of the black earth zone in the European part of the Russian Federation. For this area, data from recent (2012-2016) soil surveys as well as archived information (from the 1970s) with an average point density of at least 1 point per kilometer and large-scale soil maps at a scale of 1:25000 - 1:50000 were used.

Used: more than 150,000 point observations (during 2012-2016) and more than 15,000 polygons.

Maps a) and b) were combined into a single layer of soil organic C stocks by calculating the grid values of the given step for each of the maps and then superimposing more detailed maps b) over the map a).

An error map is also provided for this combined layer.

As the carbon stock map was developed using different approaches and methods, different approaches were also used to map errors. For most of Russia, the relative carbon stock error in the layers below the litter is estimated for each legendary margin. The maximum error is noted for map legend margins, where volumetric density and stocks are defined as the average of single, significantly different profiles. The relative error exceeds 200% in some cases.

For the high-density black earth area, the relative error in carbon stock estimation is about 25%.

For some regions, regression equations were used to calculate the volumetric density, in which case the standard deviation is about 25%.

Errors for most of the territory were estimated as partial from the division of the standard deviation by an average value. The normal distribution was assumed.

Soil allocation names, minimum, maximum and average values of the volumetric masses estimation were prepared in the form of Excel sheets and used as related tables in ArcGIS calculations.

A separate layer of organic C stocks was also formed in the humus and peated horizons of semi-hydromorphic soils. The lack of empirical data characterizing these horizons did not allow for a correct assessment of their variability, and therefore this layer was not taken into account in the error estimation map.

c) The map of carbon stocks in the litter was prepared on the basis of the previously published one (D.G. Shchepashchenko, L.V. Mukhortova, A.Z. Shvidenko, E.F. Vedrova, organic carbon stocks in soils of Russia // Soil science. 2013. № 2. С. 123-132) map adapted to GSOC17 requirements. This layer was not taken into account in the error assessment map.

The world map GSOC17 includes the Russian map in the form of the sum of all three layers (mineral part and organic - litter and peated). All maps are presented in the .tiff (float 32 bit) format in WGS84 (4326) GCS with ½ arc minute pitch within -180:180 degrees. The values are measured in kilograms C per square meter.

All initial (vector and raster) map layers in the form of ArcGIS v.10.1 project are archived and available by reference.

Summing up the first results of the work, we estimate the total organic C reserves in the 30 cm soil layer of the Russian Federation at 151Pg (Gt) (according to FAO calculations, taking into account the mismatch between water body boundaries and administrative boundaries - 149) Pg (Gt).

Almost half of these reserves (45%) are in organogenic horizons - peat layer of peat boggy soils, peat and humus horizons of semi-hydromorphic soils and litter.

Appendix

For many soil sections there is no information about the volume mass of their horizons. On the basis of the information accumulated in the IS SGDR, the applicability of the equations proposed by O.V. Chestnykh and D.G. Zamolodchikov (Dependence of soil horizons density on the depth of their occurrence and humus content) was assessed. Soil science 2004, No.8, p.937-944) for determination of volume mass of different soils,

BW = a₁ – a₂/(MID+a₃) + a₄/(HUM+a₅), where:

• BW — volume weight, g/sm³
• HUM — humus content in the horizon, %
• MID — average horizon depth, sm

Parameters of the equation

It has been revealed that for "taiga" soils ("podzolic" in the terminology of O.V. Chestnykh and D.G. Zamolodchikov) - podzolic, podzolic gley and gley, sod-podzolic, gray forest, podzolic, podbur, brown forest soils, etc. - and "humus accumulative" - all the legends of chernozems and chestnut soils, including meadow, ashy and saline soils - the values of mineral horizons density calculated using the proposed equations are quite well coincide with the experimental values. Mineral horizons with organic matter content less than 15% were referred to as mineral horizons.

For soils of the two groups of soils, the density of the horizons (in the absence of direct definitions) was determined on the basis of these equations. Then the organic C reserves in the mineral horizons of soils were calculated up to a depth of 30 cm, and then they were averaged for each of the legend's distinctions.

For the remaining mineral soils, density parameters are derived from average statistical data or expert judgement.

C stocks in swampy peatlands (upland, transition and lowland) soils were estimated taking into account the ash content and density of the corresponding peat based on the characteristics given in the survey works:

• Efremov S.P., Efremova T.T., Melentyeva N.V. Carbon reserves in the bog ecosystems // Carbon in the forest and bog ecosystems of Russia. Krasnoyarsk. 1994. С. 128-140;
• Vompersky S.E., Ivanov A.I., Tsyganova O.P., Valyaeva N.A., Glukhova T.V., Dubinin A.I., Glukhov A.I., Markelova L.G. Waterlogged organogenic soils and swamps of Russia and carbon stock in their peat / Soil science. 1994. № 12., С. 17—25;
• Inisheva L.I. A., Smirnov O. N. Carbon deposit and emission by bogs of Western Siberia Scientific dialogue. 2012 Issue No. 7. EFFECTIVE AND ECOLOGIES I p. 61—74.