Kazakhstan’s soil properties have yet to be comprehensively characterized. We sampled 40 sites consisting of ten major soil types at spring (wet) and late-summer (dry) seasons. The sample locations range from semi-arid to arid with an annual mean air temperature from 1.2 to 10.7 °C and annual precipitation from less than 200 to around 400 mm. Overall topsoil total (STC), organic (SOC), and inorganic (SIC) carbon did not change significantly between spring and late summer. STC and SOC show a wave like pattern from north to south with two maxima in northern and southern Kazakhstan and one minimum in central Kazakhstan. With a few exceptions SIC content at northern sites is generally low, whereas at Lake Balkhash SIC can exceed 75% of STC. Independent of the seasons, SOC significantly differed among soil types. Total nitrogen content distribution among our sampling sites followed a similar pattern as SOC with significant differences between soil types occurring in northern, central and southern Kazakhstan.
Machine-accessible metadata file describing the reported data (ISA-tab format)
Background & Summary
Kazakhstan is the largest land-locked country in the world. Its massive land area of 2.725 × 106 km2 represents a key reservoir for soil organic carbon that is thought to play an important role in global climate-carbon modelling1,2. Yet SOC data pertaining to Central Asia3 and particularly Kazakhstan4 are scarce, resulting in considerable ignorance of the extent to which forests and steppe soils in the north and east, and arid areas in the south and west, contribute to sequestering atmospheric carbon or yielding carbon into the atmosphere. Modelling approaches, such as the arid ecosystem model (AEM), yielded high soil organic carbon densities of up to 34 Pg for top soils of 1 m depth and 12–14 Pg for 30 cm deep top soils in temperate deserts of Central Asia5. Sub-regional SOC studies utilizing satellite imagery of farm land in northern Kazakhstan and temporal carbon variations lacked predictive power6. Normalized difference vegetation index (NDVI) values that reflect vegetative sensitivity to climate change were found to be high in northern Kazakhstan, and in central Kazakhstan, and NDVI changes were positively correlated with the annual temperature7. Net ecosystem CO2 exchange (NEE) studies of two sites representing alkaline desert soils near Lake Balkhash and the Aral Sea revealed CO2 flux dependence on moisture, pH and light8. The authors reported net CO2 release nocturnally and on cloudy days with precipitation, whereas on sunny and dry days, CO2 was taken up. Overall, the current picture of carbon cycling in Kazakhstan is supported by only a few sites and little data. In addition numerous SOC centric studies9,
More data and improved coverage of fundamental soil properties including pH, moisture content, particle size, and cation and anion composition that influence microbial activity, and thus the rate of SOC decomposition, are needed to synthesize better agro-economic and ecological strategies, especially regarding climate change predictions. Towards this end we assessed the current soil properties of Kazakhstan by sampling top soil (15 cm) at 40 sites between Petropavlovsk (north) and Taraz (south) during wet and dry seasons and determined the physical and chemical compositions, related vegetation, land cover and climate properties.
The study has 40 sampling sites that were located a minimum of 50 m from the nearest road. Sample sites were approximately 50 km apart with flat topographic conditions (toeslope) (Fig. 1). No permits were required for the sampling site locations and the sites did not harbour endangered or protected species. Geographical coordinates (WGS-84) were recorded using a Garmin T650 hand-held global positioning system. The coordinates and geographical annotations are shown in Table 1 (available online only). Sampling was conducted in 2015: one in the “wet” season after the snowmelt (May), and the other in the “dry” season at the end of the growing season (September).
Sampling site documentation
Pictures of the landscape and vegetation were taken at each location (see Supplement 1). The soil sampling procedure was divided into physical properties sampling, chemical sampling, and biome sampling. Soil sample preparation conditions for chemical and physical analyses (including depth of cores, particle size and milling) were chosen to be harmonized with future analyses of soil microbiomes15,16. Samples for biome (not subject of this paper), physical and chemical analysis parts (see Laboratory Methods) were transported on ice and transferred to 4 °C refrigerators for extended storage.
Samples for physical analyses
Samples for gravimetric moisture determination were obtained by digging a 10 cm diameter hole with a spade. The spade was rinsed well with distilled water between samples. Approximately 1 kg of soil was excavated, mixed and sealed in a 1 L plastic bag. For bulk density determination, a 50 ml conical tube (Corning Inc.) was filled with undisturbed soil derived from 15 cm deep cores and weighed on a portable balance (Maxx-412, Denver Instrument). At five sample sites additional samples for bulk density measurements were obtained to verify the cone method against the traditional soil ring method.
Samples for chemical analyses
After removing the litter layer if present, (approximately top 2 cm) soil cores of 1.5 cm in diameter were taken to depth of approximately 15 cm. The cores were transferred into 50 ml falcon tubes that were sealed with Parafilm M (Bemis Company, Inc.).
Soil moisture and bulk density
Soil moisture was calculated gravimetrically as wet and dry soil weight ratios. Briefly, triplicates of 10 g of soil were placed on aluminium paper, weighed and transferred to a 105 °C ventilated oven (Heraeus LUT 6050, Therma Fisher Scientific Inc.). After 24 h, samples were removed, immediately weighed and the dry mass recorded. Volumetric soil moisture content was calculated based on oven-dry bulk densities assuming 15 cm soil depth and by subtracting the mass of evaporated water from the wet bulk density. The volumetric soil moisture and calculated carbon stock values represent averages from two oven-dried bulk densities values obtained from two independent samples.
Soil milling for pH, conductivity and elemental analysis
Ten grams of dried and sieved soil of smaller than 150 μm particle size15 was transferred into the jar of a vibrating ball mill (MM 400, Retsch GMBH) for 2 min milling at 30 Hz. The milled sample was collected in a 15 ml Falcon tube. The vibrating ball mill processing reduces the sample particle size to approximately 1 μm.
Soil pH and conductivity
Soil pH and electrical conductivity were measured in triplicate in aqueous solution suspension (supernatant) with a 1: 2.5 (soil : water) ratio according the protocol of Pansu and Gautheyrou (2006) using a 8107UWMD Ross Ultra pH/ATC triode (Thermo Fisher Scientific, USA) and Orion 013010MD conductivity cell (Thermo Fisher Scientific, USA) electrodes.
Total soil carbon (SC), organic (SOC) and inorganic carbon (SIC) based on elemental analysis
Total SC and SOC were measured using a CNHS-O dry combustion elemental analyzer Multi N/Cb 3100, (Analytik Jena, Germany). For total SC 100 mg milled soil was placed in a ceramic combustion boat and combusted under pure oxygen at a flow rate of 2.8 L/min at 950 °C. The CO2 emitted during combustion is detected by a thermal conductivity detector. Milled soil samples for SOC estimation were pre-treated in combustion boats with 100 μl of H3PO4 (30–40%) to dissolve carbonates. Samples were dried overnight at 70 °C and subjected to combustion at 950 °C under 14 L/min oxygen (Multi N/Cb 3100, (Analytik Jena, Germany)). SIC was calculated as the difference between total SC and SOC (Fig. 2 and Table 2).
Loss-on-ignition (LOI) procedure for soil organic matter (SOM)
The soil samples were dried and sieved through a 2 mm sieve. We adapted the LOI method described by Emmett et al. (2007; Countryside Survey: Soils Report from 2007). In brief, crucibles were washed and then rinsed three times with distilled water, dried for 40 min at 105 °C in an oven (Carbolite PN 60) and then cooled to room temperature (RT) in a desiccator for 30 min. Each crucible weight was recorded (Wc) and 10.00 g of crushed soil sample (<2 mm) was weighed in the crucible, dried for four hours at 105 °C and cooled to RT in a desiccator for 30 min and the dry sample weight (Ws) recorded. Dried samples were loaded into a muffle furnace (Carbolite ELF 11/6B), heated to 375 °C for 16 h and allowed to cool down to 150 °C before being transferred to a desiccator for 30 min to cool to RT. The weight of the samples was recorded as Wa and LOI was calculated as (Ws − Wa)/(Ws − Wc) × 100.
Total nitrogen (TN) measurement by elemental analysis
The total nitrogen content of each sample was analyzed by quantitative combustion in excess oxygen using DuMaster D-480 analyzer (Büchi Labortechnik AG). L-glutamic acid was used for calibration (N-factor) of the sample measurement series. Dried and milled samples were weighed in portions of 700 mg, packed in tin foil and loaded on the sample carousel for total nitrogen measurement according to the manufacturer’s instructions (Fig. 3).
The Climate Research Unit high-resolution dataset (CRU TS v. 3.24.01 https://crudata.uea.ac.uk; January 26, 2017 release) contains air temperature and precipitation data ranging from 1901 until 2015 at 0.5° resolution of grid-boxes17 Google Earth Interface Pro was used to download the raw monthly temperature and precipitation data for the sample locations. If multiple sample locations (1–3, 4–5, 32–33, 28–29, and 19–20) mapped to one grid-box, the same temperature and precipitation data were used (e.g., sample sites 1–3 were assigned the same climate data). The climate data are presented in Table 1 (available online only).
Carbon and nitrogen stocks calculation
To infer soil total organic carbon (TOC) and nitrogen (TN) stocks in tons of TOC and TN per hectare (tC/ha, tN/ha) we followed the procedure outlined by Rowell (1994)18, (section 3.7, p. 55). The soil layer was assumed to represent the 0–15 cm depth. We used oven dry bulk densities averages from the two samplings (May and September). For TOC and TN absolute content (g/kg or mg/g) we used mean values of averaged duplicate measurements (elemental analysis) obtained from two samplings (May and September) see Fig. 4.
The data were deposited at the Mendeley data repository (Data Citation 1: Mendeley Data https://doi.org/10.17632/5ZMVX4VH28.1) as Supplement 1 and 2. Supplement 1 contains a folder with photos of landscape and vegetation for each sample location. The naming convention is location number. photo number (date the sample was taken). For example, 1.1(25.05.15) corresponds to location 1 and photo 1 on May 25, 2015. For the images presented in our data records which feature identifiable human participant(s) the informed consent was obtained from the participants prior to publication of the images. Supplement 2 is a Microsoft Excel file that contains spreadsheets with data on 1) soil TC, soil total organic carbon (TOC), soil total inorganic carbon (TIC), 2) Loss on Ignition (LOI), 3) soil TN, 4) soil dry bulk density, 5) gravimetric and volumetric soil moisture 6) soil suspension and supernatant pH, 7) soil electrical conductivity, and 8) soil TOC and TN 9) annotation table (Table 1 (available online only)). Values of sample measurements from each location were reported with standard deviation (STDEV) and standard error (SE) when applicable.
Soil carbon and nitrogen concentrations were averaged from duplicate sample preparations and validated using reference standards. If the concentrations deviated significantly the measurements were repeated. SOC results were validated independently by conducting LOI measurements (see Fig. 5). Soil organic matter (SOM) (g/kg) data obtained from LOI analyses were converted to TOC (g/kg) by multiplying the values with the coefficient 0.5819.
How to cite this article: Yapiyev, V. et al. Top soil physical and chemical properties in Kazakhstan across a north-south gradient. Sci. Data. 5:180242 doi: 10.1038/sdata.2018.242 (2018).
Publisher’s note: Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Yapiyev, V. et al. Mendeley Data https://doi.org/10.17632/5ZMVX4VH28.1 (2018)
This research was supported in part by the grant “Kazakhstan soil microbiome: agricultural characteristics and perspectives” in the framework of the agreement #104 dated 12.02.2015 of the program #055 “Scientific and/or Scientific-Technical Activities” within subprogram #101 “Research Grant Funding” from the Ministry of Education and Science of the Republic of Kazakhstan and by Nazarbayev University. We thank Jong Kim for letting us use the equipment of Department of Civil and Environmental Engineering, Almagul Kushugulova and Samat Kozhakhmetov for administrative and managerial support and Nurlan Ongdas for helping to collect field samples. We are grateful for the valuable input and assistance provided by Drs. Joanna Clark, Tom Sizmur, Maria Shahgedanova, and Anne Verhoef at University of Reading.