Soil carbon stock, litter decomposition, and weather data from Ethiopian forests

Introduction

100 sampling units (SU) were selected from the total of 631 SUs of the Forest Reference Level submission 2017 (FRL 2017). The sampling was designed unbiased for total growing stock per SU, altitude,and mean litter depth per SU. The actual field sampling succeeded on 98 of the pre-selected SUs due to accessibility restrictions.

Soil profile sampling

Soil sampling was performed from November 2017 till mid-January 2018. Samples were taken from undisturbed soil from depths of 0-10 cm, 10-20 cm, and 20-30 cm below the organic layer. Volumetric samples of 107.5 cm3 were taken vertically, using a 10 cm long conically shaped corer with a cutting lower edge diameter of 37 mm and upper diameter of 40 mm. 

Composite samples were formed by combining the volumetric samples taken from different depths of two parallel soil profiles. The samples were transported to EEFRI Soil Laboratory in Addis Ababa after 1-4 weeks of sampling at distant locations. 

Soil physical characteristics

The soil samples were air-dried, homogenized, and subjected to oven-drying at 105°C until constant mass. Total bulk density was determined using the total dry mass and volume of the composite samples. 

Organic carbon content (C % by wet oxidation method), and soil physical characteristics: moisture content, bulk density of the total sample, and bulk density of fine fraction (particles passing the 2 mm sieve). The mass of the coarse fraction was weighed. The soil fine fraction was also subjected to laser diffraction for more accurate particle size analysis for proportions of clay, silt, and sand. 

In addition to this 28 samples were also analyzed for C content in the laboratory of Natural Resources Institute Finland to determine C content by LECO CHN analyzer. This was done to calibrate the bulk of wet digestion-based estimates (Fig. 1). Before analysis, the soils were tested for the presence of inorganic C.

For Figure 1. See Soil_C_Ethiopia.pdf

Figure 1. Comparison of results from wet oxidation (Walkley-Black) and dry oxidation (CHN analyzer). The dotted line shows the theoretical 1:1 match between the axis, the solid line shows linear regression (intercept = 0) between the methods. The estimated slope value of 1.165 was used in adjusting the wet digestion results to match those obtained by dry oxidation: OCadj = 1.165 * OCwet.

Based on a linear regression between the wet and dry oxidation analysis results, a correction factor of 1.165 was applied to adjust the organic C% obtained by wet digestion. The adjusted data are shown in the file “SOC_Ethiopia_2017-2018.csv”.

SOC stocks were calculated by multiplying the proportion of organic C with BD of fine earth, after which the result was corrected for stoniness, a visually estimated proportion of large stones (S, value from 0 to 1) in the soil profile that could not be included in the volumetric soil samples (FAO VS-FAST).

\(SOCstock = C_{org} * BD_{fe} * (1-S)\)

Soil organic carbon stock data

Files: “SOC_Ethiopia_2017-2018.csv” and “SOC_Ethiopia_2017-2018.xlsx”

The file includes soil characteristics from layers of 0-10 cm, 10-20 cm, and 20-30 cm below the loose organic layer on top of the soil. The data are used for SOC stock estimation in the respective layers as described above.

In the .csv file individual columns are for 

LAT is the latitude of the sampling site corresponding to FieldCode and SU_nr

LON is the longitude of the sampling site corresponding to FieldCode and SU_nr

The coordinates are expressed as decimal degrees of the WGS84 system

FieldCode refers to the Region and Sampling Unit number of the Ethiopian NFI (see below) 

SU_nr is the Sampling Unit number of the Ethiopian NFI

Region is the name of the administrative region where the sample was taken

Biome is the name of the forest biome type where the sample was collected

BiomeSimplified is the name of a biome with some close types combined

DepthRange is the upper and lower limit of the soil sample in the field, cm

StoninessVFAST is a percentage of stones (VS-FAST by FAO) in the ca. 40 cm deep soil profile exposed during the sampling

FreshMassInField is the mass of the total composite soil sample of the given layer, g, primarily indicative of checking the correct number of subsamples in composite

NrComposites is the number of subsamples included in the composite for each soil layer

CorerVolume is a constant of 107.5 cm3 because only one type of corer was used for undisturbed, volumetric sampling

CompositeVolume is the volume of the composite sample for each soil depth layer

CoarseFractionMass is the dry mass, g of soil particles > 2mm that did not pass the sieve, but were included in the sample volume

FE_DryMass is oven-dry mass, g of the fine fraction that passed the 2 mm sieve.

BDtot is total bulk density, g m-3, calculated for the composite sample

BDfe is the bulk density of the fine earth fraction, g m-3

OC_adj is organic carbon (OC) content (%) in the composite sample, adjusted according to the comparison between dry and wet oxidation methods (Fig. 1)

SOCfe is SOC stock calculated for soil fine earth fraction, t ha-1 in the 10 cm deep soil layer

SOCfe_stoniness is SOC stock of the fine earth fraction, t ha-1 in the 10 cm deep soil layer, adjusted for stoniness. The correction assumes that the volume occupied by larger stones would be void of OC. 

Litter stock data

File: “Litter_Ethiopia_2017-2018.csv”

The file includes measurements of litter layer on Ethiopian NFI Sampling Unit (SU) sites where sampling for SOC stock determination was done. The depth of the litter layer was measured in the SU’s of the NFI, and this data contains in addition to depth also a volumetric sample of the litter layer. The dry bulk density was used to calculate the carbon stocks in the litter pool.

The depth of the litter layer was measured in the field. Litter from the respective spot was sampled quantitatively from a frame of 0.01m2 of area for litter dry mass estimate.

The organic C stock in a litter (L) was calculated as,

\(L = {M\over z} * {C_{om}\over A}, \)

where

M = Dry mass of the litter sample, g

z = Depth of the litter layer in the field, m

Com = Conversion factor from dry organic matter to carbon (C), 0.5

A = area of quantitative collection of litter (0.01 m2)

In the .csv file individual columns are for

LAT, LON is the GPS coordinates (decimal degrees of WGS84) for the Sampling Units (SU_ID)

SU_ID is the Sampling Unit identification number of the Ethiopian NFI

FieldCode refers to the Region and Sampling Unit number of the Ethiopian NFI (see below)

Region is the name of the administrative region where the sample was taken

Litter_dry is the dry mass, g of the litter sample

Area_m2 is the area, m2 of litter sampling

MeanLitterDepth is the mean depth of the litter layer at the sampling area

CDensityLitter is the dry bulk density of the litter, g m-2 multiplied by the assumed organic C proportion of the oven-dry litter materials (0.50)

LitterCStock_tha is the litter stock, t ha-1 calculated from the C density of the litter layer

Litter bag data (decomposition and quality)

The leaves and twigs were sampled from 2 species (Juniperus and Podocarpus) and 3 locations of the elevation gradient in the Chilimo forest (Table 1). The forest was considered an old-growth with Juniperus procera and Podocarpus falcatusbeing the main species forming the tree canopy. The sites form an elevation gradient (Table 1).

Table 1. Geographical locations of the study sites in the Chilimo forest.

id

Latitude (deg.)

Longitude (deg.)

Elevation

(m a.s.l)

1

9.0672

38.1443

2500

2

9.0712

38.1556

2670

3

9.0869

38.1684

2800

The dying and dead leaves were sampled directly from the trees later referred to as “fresh” and from the branches found on the ground, referred to as “old”. The old leaves were assumed to be dead for around 3 months. The diameter of the branches/twigs was less than 1 cm in diameter. The samples were first sorted and air-dried in an elevated temperature of the greenhouse and thereafter oven-dried in the oven overnight at 45 °C.  The samples were analyzed for acid, water, ethanol dissolved,and undissolved fractions (AWEN) (Table 2) and for the decomposition rates of the litter installed into the litter bags corresponding to each of the Chilimo sites. 

Table 2. Acid, water, ethanol (A, W, E, respectively) dissolved and undissolved fractions (N) from the litter components of the dominant tree species in the Chilimo forest.

Litter type

Species

A

W

E

N

leaves fresh

Juniperus 

0.45

0.13

0.1

0.33

leaves fresh

Podocarpus 

0.42

0.28

0.05

0.25

leaves old

Juniperus 

0.44

0.07

0.08

0.41

leaves old

Podocarpus 

0.44

0.09

0.05

0.42

twigs

Juniperus 

0.61

0.04

0.02

0.32

twigs

Podocarpus 

0.56

0.15

0.02

0.27

A sufficient amount of litter was placed into the litter bags (polyurethane mesh 1 mm) and the mesh bags were installed on top of the soil surface under the forest canopy (later referred to as “canopy”) and in the forest gap caused by harvesting (later referred as “open”). The installation of the litter bags (for each species 3 replicates of each litter type for each site and canopy type for the 3 periods, in total 12 litter bags for leaves and 6 bags for twigs) was done on 22.9.2017. The mesh bags were left on the ground, protected from grazing by the fence, and retrieved subsequently on 12.10.2017, 31.10.2017, and 12.12.2017. Despite the efforts took few samples were lost. The retrieved samples were oven-dried and initial mass and mass loss data for each period and litter type with a detailed description of the variables can be found in the file “litter.chilimo_07.02.22.xlsx”.

Soil temperature data

During the period from 22.9.2017 to 12.12.2017, we monitored the soil temperature at 5 cm depth under the canopy and in the open canopy on all Chilimo sites continuously every 4 hours intervals with the Maxim iButton temperature loggers. However, some sensors were lost. Daily means and their standard deviation of the continuous temperatures can be found in the file “soil.temp.chilimo_07.02.22.xlsx”.

Processed weather data

The air temperature and precipitation data for 98 sampling units corresponding to soil carbon data originated from 73 weather stations located across Ethiopia and were obtained from Ethiopian Meteorological Agency (http://www.ethiomet.gov.et/). Sampling units were joined with weather data by the closest proximity to their corresponding weather stations. Precipitation was unaltered. The air temperature required correction by elevation is described in more detail in Lehtonen et al. (2020). The monthly values of air temperature and precipitation with an accompanied readme description of the variables can be found for 98 sampling units in the file “sampling.units98_meteo_07.02.22.xlsx” and the Chilimo study sites in the file “monthly.weather.chilimo_07.02.22.xlsx”. The monthly values in the file "sampling.units98_meteo_07.02.22.xlsx" correspond to long-term average over the period from 1986 to 2017.

References:

Lehtonen, A., Ťupek, B., Nieminen, T.M., Balázs, A., Anjulo, A., Teshome, M., Tiruneh, Y. and Alm, J., 2020. Soil carbon stocks in Ethiopian forests and estimations of their future development under different forest use scenarios. Land Degradation & Development, 31(18), pp.2763-2774.

FRL 2017. https://redd.unfccc.int/files/ethiopia_frel_3.2_final_modified_submission.pdf