1 Introduction
2 Materials and methods
2.1 Study area
2.2 Sampling set
Fig. 1 Distribution of sampling sites in four ecosystem types (Farmland, 19; Grassland, 22; Forestland, 28; Shrubland, 28) across the Loess Plateau. Digital elevation models (DEMs) represent the elevation across the Loess Plateau. DEM data were downloaded from the United States Geological Survey and are free for public use. The figure was generated with ArcGIS 10.0 (http://www.esri.com/) |
2.3 Estimation of plant C stocks
Table 1 Biomass models and parameters of different trees. Note: WS meas biomass of trunk; WP meas biomass of stem; WB meas biomass of branches; WL meas biomass of leaves; WT meas total above-ground biomass; WR meas total below-ground biomass |
Tree species | Biomass models and parameters |
---|---|
Picea asperata | WT = 0.067732(D2H)0.865949; WR = 0.0088D2.53827 |
Hemlock | WT = 0.149707(D2H)0.80139; WR = 0.19758D0.6058 |
Larix gmelini | WT = 0.046238(D2H)0.905002; WR = WT/4.81 |
Pinus tabuliformis | WS = 0.027636(D2H)0.9905; WB = 0.0091313(D2H)0.982 WL = 0.0045755(D2H)0.9894; WT = WS+ WB+ WL; WR = 0.0084800D0.988 |
Pinus armandi | WS = 0.01308(D2H)1.0038; WB = 0.0055(D2H)1.0439 WL = 0.0011(D2H)1.12566; WT = WS+ WB+ WL; WR = 0.0033D1.0148 |
Pinus massoniana | WT = 0.071556(D2H)0.857209; WR = WT/6.23 |
Cunninghamia | WS = 0.073429(D2H)0.86262; WP = 0.013775(D2H)0.84463; WB = 0.000482(D2H)1.23314 WL = 0.019638(D2H)0.78969; WT = WS+ WP |
Cupressus | WS = 0.12531(D2H)0.733; WB = 0.137403 + 0.012887D2H WL = 0.05349 + 0.00997D2H; WT = WS+ WB+ WL; WR = 0.01109 D2H-0.160386 |
Betula | WT = 0.0278601(D2H)0.993386; WR = WT/2.89 |
Broad-leaved forest | WS = 0.044(D2H)0.9169; WP = 0.023(D2H)0.7115; WB = 0.0104(D2H)0.9994; WL = 0.0188(D2H)0.8024; WT = WS+ WP+ WB+ WL; WR = 0.0197D0.8963 |
Coniferous forest | WT = 0.0495502(D2H)0.952453; WR = WT/3.85 |
2.4 Estimation of soil C stocks
2.5 Estimation of ecosystem C stocks
2.6 Climate factors and soil properties
2.7 Data analysis
3 Results
3.1 Spatial distribution of C densities across the loess plateau
Fig. 2 Spatial pattern of ecosystem total carbon density (ETCD) along longitude and latitude across the Loess Plateau |
Fig. 3 Spatial distribution of above-ground carbon density (AGCD) (a), below-ground carbon density (BGCD) (b), soil organic carbon density (SOCD) in 0-10 cm layer (c), and soil organic carbon density (SOCD) in 10-20 cm layer (d) across the Loess Plateau |
3.2 Carbon stocks across the loess plateau
Fig. 4 Summary of the ecosystem C stocks across the Loss Plateau of China (a). Statistics of C density and stocks in the Loess Plateau and China derived from various studies, compared with our study (b). AGCs: above-ground carbon stock; BGCs: below-ground carbon stock; SOCs: soil organic carbon stock; ETCs: ecosystem total carbon stock |
3.3 Carbon densities across ecosystems and climate zones
Fig. 5 Box plots show the C density in above-ground, below-ground and soil in farmland, grassland, shrubland, forestland. The lower and upper boundaries of the box represent the first and third quartiles, respectively; the horizontal line represents the mean; the bounds of the lower and upper bars reflect the 10th and 90th percentiles, respectively. The proportion of above-ground carbon density (AGCD), below-ground carbon density (BGCD) and soil organic carbon density (SOCD) in the ecosystem total carbon density (ETCD) in farmland, grassland, shrubland, forestland Average mean annual temperature (MAT) and precipitation (MAP) impacted ecosystem C densities depending on ecosystem types (Fig. S3). The SOCD was obviously higher in zones with temperature higher than 10 °C. The AGCD and BGCD, ETCD in zones with temperature lower than 5 °C were higher than in warmer climate. The SOCD had the same trend as those of AGCD and BGCD. Thus, the ETCD was obviously higher in zones with temperature lower than 5 °C, compared with those with the temperature higher than 10 °C (p < 0.05). Similarly, SOCD, AGCD and ETCD showed the same change trend, and increased within zones with precipitation higher than 500 mm, although BGCD increased within precipitation lower than 250 mm. Thus, wetter and warmer conditions contributed to superior biomass productivity and so, to ecosystem C accumulation. |
3.4 Effects of climate conditions and soil properties on C densities
Fig. 6 Pearson correlation between C density and climate factors and soil properties for four ecosystem types. The color of each roundness is proportional to the value of Pearson’s correlation coefficient. Green indicates a positive correlation (dark green, r = 0.80); orange indicates a negative correlation (dark orange, r = 0.80). * p < 0.05; ** p < 0.01. AGCD: above-ground carbon density; BGCD: above-ground carbon density; SOCD: soil organic carbon density; ETCD: ecosystem total carbon density; MAP: mean annual precipitation; MAT: mean annual temperature; pH: soil pH; TN: soil total nitrogen; CEC: soil cation exchange capacity |
Fig. 7 The variation partitioning analysis of the relative explanatory rate of environmental factors to explain AGCD, BGCD, SOCD, ETCD. The circular rings mean the relative explanatory rate of soil and climate factors to AGCD, BGCD, SOCD, ETCD. The explained variability was calculated after 999 bootstraps. AGCD: above-ground carbon density; BGCD: above-ground carbon density; SOCD: soil organic carbon density; ETCD: ecosystem total carbon density; MAP: mean annual precipitation; MAT: mean annual temperature; pH: soil pH; TN: soil total nitrogen; CEC: soil cation exchange capacity |
Fig. 8 Structural equation models (SEMs) depicting the multiple relations of ecosystem total carbon density (ETCD) with climate factors and soil properties, and the values are standardized coefficients of the models. The solid orange lines are the negative relationships, and the solid green lines are the positive relationships, and the solid black lines are no significant relationships, respectively. Arrows represent a directional influence of one variable upon another. The numbers beside the arrows are standardized coefficients. The thickness of the arrows is proportional to the magnitude of the standardized path coefficients or covariation coefficients. R2 stands for variation interpreted by variables, which is calculated after 999 bootstraps, and the significant level is set at α = 0.05, *p < 0.05, **p < 0.01. AGCD: above-ground carbon density; BGCD: above-ground carbon density; SOCD: soil organic carbon density; ETCD: ecosystem total carbon density; MAP: mean annual precipitation; MAT: mean annual temperature; pH: soil pH; TN: soil total nitrogen; CEC: soil cation exchange capacity |