Amount, composition, and turnover of organic matter pools in grassland soils under typical steppe vegetation types of the Xilin River Basin (Inner Mongolia) with different grazing management
Zusammenfassung der Projektergebnisse
Summary of results (1st and 2nd project phase; 3-6 pages) First project phase: In the first phase, five sites with different grazing intensities were selected in Leymus chinensis dominated areas (ungrazed since 1979 = Ug79, ungrazed since 1999 = Ug99, winter grazing = Wg, continuously grazed = Cg, heavily grazed = Hg). For the statistical and the geostatistical approach topsoils (0-4 cm) of two differently sized regular, orthogonal grids (small grids with 100 sampling points: 15 m spacing, 5 m nested sampling; large grids with 125 sampling points: 50 m spacing, 10 m nested sampling) were sampled. Differently sized grids allowed the exploration of scale effects. Each sample was analysed for bulk density, organic carbon (OC), total N and total S concentration, δ13C, pH, Ah horizon thickness, vegetation cover and aboveground biomass. The dataset was analysed using general statistics, multivariate explorative statistics, variograms and cross variograms. In each of the five plots representative soil pits were sampled to analyse effects of grazing and grazing cessation on the amount, composition and turnover of SOM in detail. A combined aggregate size, density and particle size fractionation procedure was applied in three horizons of each pit to separate functional SOM fractions and pools. Additionally aggregate stability measurements were done on topsoils. Statistical analyses showed that bulk density increased significantly with increasing grazing intensity. OC, total N and total S concentrations decreased significantly with increasing grazing intensity. No effect on the pH or C/N ratio was detected. Significant differences in C/S and N/S ratios between differently grazed plots were found. These differences pointed towards a relative accumulation of sulphur in grazed compared to ungrazed areas following an increased organic matter decline or lower inputs of diluting litter. Elemental stocks of the upper 4 cm were calculated for OC, total N and total S using the measured bulk densities. The data revealed significantly lower amounts for all three elements on the heavily grazed site, but no significant differences for the other areas. In addition, elemental stocks were calculated using an equivalent mass instead of bulk density to take into account changes in P1 - Amount, composition, and turnover of organic matter pools in grassland soils under typical steppe vegetation types of the Xilin River Basin (Inner Mongolia) with different grazing management. 76 bulk density following grazing. This revealed a highly significant decrease for OC, total N and total S with increasing grazing intensity. OC, total N and total S concentrations respond similarly to different grazing intensities, showing highly significant positive correlations. OC concentrations and bulk densities were significantly negatively correlated. Effects of grazing cessation were only found in the long-term, while no ameliorating effects of reduced or excluded grazing could be detected five years after grazing cessation. After 25 years of exclusion, significantly different values were found for all parameters. The statistical approach showed that physical and chemical parameters of steppe topsoils deteriorated significantly following heavy grazing, remained stable if grazing was reduced or excluded for five years and recovered significantly after 25 years of grazing exclusion. Geostatistical analyses showed that the spatial distribution in small grids changed with grazing intensity. Generally, heterogeneity of topsoil properties increased with decreasing grazing intensity from a homogeneous to a patchy distribution. This is attributed to vegetation recovery/succession and deposition of windblown material in ungrazed areas. Ug99 showed different spatial dependencies than continuously and heavily grazed, but has not yet reached the high variability of Ug79. Large grid sampling did not detect small-scale variability or grazing impacts, but showed spatial dependencies that were attributed to topography or soil erosion/deposition. Low OC concentration and low Ah thickness were associated with hilltop and shoulder positions, resulting in lower OC stocks at these topographic units. The geostatistical approach showed that recovering vegetation and higher deposition of windblown material around recovering plants are crucial processes that initiate the recovery of grazing-degraded areas. Physical fractionation of topsoils (0-10 cm) showed that greater inputs of organic matter led to larger amounts of OC in coarse aggregate size classes (ASC) and especially in particulate organic matter fractions (POM). No grazing-induced changes of SOM quantity were found in fine ASC and particle size fractions. SOM quality (solid state 13C NMR spectroscopy, neutral sugar analysis) was comparable between different grazing intensities, but ungrazed plots had slightly more decomposed SOM across all fractions. Ug79 showed generally greater radiocarbon concentrations compared with Cg. Aggregate stability, analysed as resistance to sonication, was greater in Ug79 compared with Cg. Larger litter inputs in grazing exclosures increased POM quantity, led to faster SOM turnover and resulted in the formation and stabilisation of coarse aggregates. Organo-mineral associations turned over faster as indicated by increased radiocarbon concentrations, but the OC content of this pool did not change. The physical fractionation of topsoils showed that additional litter inputs due to P1 - Amount, composition, and turnover of organic matter pools in grassland soils under typical steppe vegetation types of the Xilin River Basin (Inner Mongolia) with different grazing management. 77 grazing cessation were sequestered in the intermediate POM pool and the long-term pool of organo-mineral associations appears to be close to saturation. Aggregate stability and formation in topsoils was increased after grazing cessation. Higher inputs of organic matter led to higher amounts of OC in coarse ASC and especially in POM fractions across all depth. These processes started in the topsoil and took more than 5 years to reach deeper soil horizons (>10 cm). After 25 years of grazing cessation, subsoils showed clearly higher POM amounts. No grazing-induced changes of SOM quantity were found in fine ASC and particle size fractions. Current carbon loading of fine particle size fractions was similar between differently grazed plots and decreased with depth, pointing towards free sequestration capacities. Despite these free capacities, no increase in current carbon loading after 25 years of grazing exclusion could be detected. It is supposed that either the particle size fractions are already saturated and empirical estimations overestimate sequestration potentials or that the climatic conditions delay the decomposition and incorporation of OM in particle size fractions. POM quality was analysed using solid-state 13C NMR spectroscopy and was comparable between different grazing intensities. POM is decomposed hierarchically from coarse to fine particles in all soil depths and grazing cessation has not affected the OM decomposition processes. The surplus of OM due to grazing cessation was predominately sequestered in readily decomposable POM fractions across all horizons and the long-term stabilisation of OM in these steppe soils is questioned. Overall this study statistically substantiates the long process of recovery of physical and chemical soil parameters after grazing cessation. The recovery of degraded areas after grazing cessation starts with the recovery of vegetation. Larger increased litter inputs and larger accumulation of wind-blown materials around individual recovered plants act as nucleus of recovery (islands of fertility). Recovery of topsoils was detectable after 25 years of grazing cessation, but spatial distribution of vegetation showed first evidence of recovery already after 5 years. In ungrazed topsoils the increased litter inputs were predominately stored in the readily decomposable POM fractions. Abundance and stability of coarse and medium aggregates were increased in ungrazed topsoils as consequence of higher POM amounts. Particle size fractions in topsoils did not change quantitatively after grazing cessation pointing towards saturation of their carbon sequestration potential. Despite saturation higher radiocarbon concentrations of particle size fractions in ungrazed plots show that the generally as stabile assumed pool of organo-mineral associations is taking part in the carbon cycling. Deeper soil horizons received higher OM inputs after 25 years of grazing cessation. But these inputs were also stored in the readily decomposable POM fraction and P1 - Amount, composition, and turnover of organic matter pools in grassland soils under typical steppe vegetation types of the Xilin River Basin (Inner Mongolia) with different grazing management. 78 are therefore not stabilised in the long-term. Finally, grazing reduction and especially grazing cessation can help mitigating the detrimental effects of heavy grazing on semiarid steppe soils when management is controlled for periods longer than 25 years. But if grazed steppe soils can become a carbon sink in the long-term when grazing management is improved or completely stopped has to be questioned. Second project phase: For the estimation of SOM storage at small and intermediate spatial scales, topsoil samples (0 – 4 cm) were taken within small plots (2 m × 2 m) at 40 points and from large grids (120 m × 150 m) at 100 points for both Leymus chinensis and Stipa grandis dominated steppe ecosystems. To reveal the impact of grazing on SOM storage continuously grazed (CG) sites were compared to long-term ungrazed (UG79) experimental sites, which were fenced in 1979. All samples were analyzed for soil organic carbon (SOC), total nitrogen (Ntot), total sulphur (Stot), bulk density (BD), pH, Ah thickness, and carbon isotope ratios (δ13C). Geostatistics were applied to elucidate the spatial distribution of SOM both at field (120 × 150 m grid) and plant (2 × 2 m plot) scale. A combined approach of soil chemical and physical analytical methods was applied to investigate the effect of grazing and grazing exclusion on the amount and stability of soil aggregates and the associated physical protection of SOC. Topsoil samples (0 – 10 cm) from CG and UG79 sites for both steppe types were analyzed for basic soil properties and separated into free and aggregate-occluded particulate organic matter fractions (oPOM) and mineral-associated fractions. Tensile strength of macroaggregates was measured by crushing tests. Undisturbed samples as well as artificially compacted samples beyond the precompression stresses were incubated and the mineralization of SOC was measured. For the assessment of early warning indicators for beginning steppe degradation, a controlled grazing experiment was established that included ungrazed (UG), moderately grazed (MG) and heavily grazed (HG) plots. Several potential soil and vegetation parameters were sampled at all sites before the start of the experiment and again after 3 years. Topsoil samples (0 – 4 cm) were analyzed for SOC, Ntot, Stot and BD, and as vegetation parameters, aboveground net primary productivity (ANPP), tiller density (TD) and leaf area index (LAI) were determined. For the calculation of SOM stocks at the catchment scale, the spatial distribution of soil organic carbon (SOC), total carbon (Ctot), total nitrogen (Ntot) and total sulphur (Stot) stocks were modelled using a Digital Soil Mapping (DSM) approach. This approach included P1 - Amount, composition, and turnover of organic matter pools in grassland soils under typical steppe vegetation types of the Xilin River Basin (Inner Mongolia) with different grazing management. 79 Random Forest (RF) as a new modeling tool for soil properties and Classification and Regression Trees (CART) as an additional method for the analysis of variable importance. The study area was classified into six land use units and at 120 locations soil profiles were sampled by horizon down to 1 m depth and analyzed for soil texture, SOC, Ctot, Ntot, Stot, bulk density (BD) and pH. On the basis of a digital elevation model, the catchment was divided into pixels of 90 m × 90 m and for each cell, the predictor variables land use unit, Reference Soil Group (RSG), geologic unit and 12 topography-related variables were determined. Results from the spatial investigations on CG and UG79 sites showed that concentrations and stocks of SOC, Ntot, Stot were significantly lower and BD significantly higher at both CG sites. At the field scale, semivariograms of these parameters showed a heterogeneous distribution at UG79 sites and a more homogeneous distribution at CG sites, whereas nugget to sill ratios indicated a high small-scale variability. At the plant scale, semivariances of all investigated parameters were one order of magnitude higher at UG79 sites than at CG sites. The heterogeneous pattern of topsoil properties at UG79 sites can be attributed to a mosaic of vegetation patches separated by bare soil. Ranges of autocorrelation were almost congruent with spatial expansions of grass tussocks and shrubs at both steppe types. At CG sites, consumption of biomass by sheep and hoof action removed vegetation patches and led to a homogenization of chemical and physical soil properties. The spatial distribution of topsoil properties at the plant scale (< 2 m) could be used as an indicator for degradation in semi-arid grasslands. The results further show that the maintenance of heterogeneous vegetation and associated topsoil structures is essential for the accumulation of SOM in semi-arid grassland ecosystems. The investigation of the impact of grazing on the physical protection of SOM revealed a higher cumulative release of CO2-C for CG sites compared to UG79 sites for both steppe types. Considerably increased contributions of oPOM were found for UG79 compared to CG sites, but macroaggregate stabilities were lower for ungrazed sites. Artificially compacted samples showed only a slightly higher C release at CG but a considerably enhanced mineralization at UG79. The continuous trampling of grazing animals together with a lower input of organic matter probably leads to the formation of mechanically compacted stable clods, which do not provide an effective physical protection for SOC in the grazed plots. At UG79, a higher input of organic matter acting as binding agent in combination with an exclusion of animal trampling enhances the formation of soil aggregates in a hierarchical order. Thus, grazing exclusion promotes the physical protection of SOC by increasing soil P1 - Amount, composition, and turnover of organic matter pools in grassland soils under typical steppe vegetation types of the Xilin River Basin (Inner Mongolia) with different grazing management. 80 aggregation and is hence a management option to enhance the C sequestration potential of degraded steppe soils. However, the aggregate-related stabilization of SOC is highly sensitive towards high grazing pressures beyond the precompression stress due to low macroaggregate stabilities that results in large losses of SOC due to mineralization. Results from the grazing experiment for the assessment of short-term indicators showed that after three years, BD increased and SOC, Ntot, Stot, ANPP and LAI significantly decreased with increasing grazing intensity. At the start of the experiment, all investigated parameters showed comparable starting conditions between all experimental sites. These parameters can be regarded as sensitive early warning indicators for degradation of semi-arid grasslands. In contrast, TD cannot be regarded as sensitive indicator. Vegetation parameters were, however, more sensitive not only to grazing but also to temporal variation of precipitation during the period of the experiment between 2006 and 2008. Contrary, soil parameters were primarily affected by grazing and resistant against climatic variations. As the investigated soil parameters were primarily affected by grazing and resistant against climatic variations, they are more suitable for assessing desertification than vegetation parameters. The assessment of starting conditions in the study area and the application of defined grazing intensities is essential for the detection of early warning indicators for grazing-induced degradation in semi-arid environments. The estimation of SOM stocks for the Xilin River catchment showed that the highest amounts of elemental stocks are stored under marshland, steppes and mountain meadows. River-like structures of very high elemental stocks in valleys within the steppe areas are partly responsible for the high amounts of SOM for grasslands (81 – 84% of total catchment stocks). For bare soil, arable land and sand dunes, very low amounts of SOM were estimated. A decline of elemental stocks of up to 70% was calculated after conversion of steppe to arable land. Analysis of variable importance showed that the most important variables that influence SOM storage are land use, RSG and geology. Topography-related variables only explain a minor proportion of the variance. Prediction accuracy of the modeling and the generated maps was acceptable with explained variances of 42 to 62% and 66 to 75%, respectively. RF in combination with CART is recommended as a suitable tool for the spatial prediction of soil properties in regions with limited availability of data. It is further concluded that soils of semi-arid steppes in Northern China are susceptible toward cultivation and thus their suitability for agricultural use is limited. To summarize, this work showed that intensive grazing in the grasslands of Northern China leads to a considerable depletion and a spatial homogenization of SOM. The continuous P2 - Topography and stocking rate: Limitation of grassland productivity by water and nitrogen. 81 trampling of grazing animals results in the destruction of soil aggregates associated with the release of physically protected SOM that is mineralized. Patches of high SOM stocks around grass tussocks are removed and thus the ability of the heterogeneous landscape to accumulate organic matter under vegetation patches is diminished. Beginning degradation of steppes is indicated even after short periods of time by changes of SOM and vegetation parameters. Grazing exclusion is an effective management option to restore degraded grassland soils as soil aggregation is promoted and SOM starts to accumulate in a patchy pattern. At the catchment scale where also areas with lower grazing intensities are included semi-arid grasslands store considerable amounts of SOM, particularly in valleys and depressions, and can be regarded as important C sinks. A further intensification of the grazing management or an extension into undisturbed boundary areas of the steppe should be prevented.
Projektbezogene Publikationen (Auswahl)
-
(2008): Grazing effects on soil chemical and physical properties in a semiarid steppe of Inner Mongolia (P.R. China). Geoderma, 143(1-2), 63-72
Steffens, M., Kölbl, A., Totsche, K.U., Kögel-Knabner, I.
-
(2009): Alteration of soil organic matter pools and aggregation in semi-arid steppe topsoils as driven by organic matter input. European Journal of Soil Science, 60, 198-212
Steffens, M., Kölbl, A., Kögel-Knabner, I.
-
(2009): Degradation and small-scale spatial homogenization of topsoils in intensively grazed steppes of Northern China. Soil and Tillage Research, 104, 299-310
Wiesmeier, M., Steffens, M., Kölbl, A., Kögel-Knabner, I.
-
(2009): Spatial variability of topsoils and vegetation in a grazed steppe ecosystem in Inner Mongolia (P.R. China). Journal of Plant Nutrition and Soil Science, 172, 78-90
Steffens, M., Kölbl, A., Giese, M., Hoffmann, C., Totsche, K.U., Breuer, L., Kögel-Knabner, I.
-
(2010): Distribution of soil organic matter between fractions and aggregate size classes in grazed semiarid steppe soil profiles. Plant and Soil, 338 (1-2), 63-81
Steffens, M., Kölbl, A., Schörk, E., Gschrey, B., Kögel-Knabner, I.
-
(2011): Digital mapping of soil organic matter stocks in a semi-arid steppe ecosystem. Plant and Soil, 340 (1-2), 7-24
Wiesmeier, M., Barthold, F.K., Blank, F.B., Kögel-Knabner, I.
-
(2011): Grazing changes topography-controlled topsoil properties and their interaction on different spatial scales in a semi-arid grassland of Inner Mongolia, P.R. China. Plant and Soil, 340 (1-2), 35-58
Kolbl, A., Steffens, M., Wiemeier, M., Hoffmann, C., Funk, R., Kruemmelbein, J., Reszkowska, R., Zhao, Y., Peth, S., Horn, R., Giese, M., Kögel-Knabner, I.
