| Soil respiration (RS) is the second largest terrestrial carbon flux and plays an important role in ecosystem carbon cycling. Rs originates from rhizospheric respiration (RR) and heterotrophic respiration (RH), and is considered to be sensitive to climate, forest age and forest conversion. Therefore, there is substantial temporal and spatial variation in RS even within a restricted geographic region. The Liangshui National Nature Reserve contains a variety of forest types with different management regimes and thus provides an opportunity to quantify the variability of RS and identify the controlling factors of the variability of RS at various temporal and spatial scales within a restricted geographic region. These findings enable us to quantify the carbon source/sink strength of ecosystems.In the present study, the ’space-for-time’ substitution was used for understanding the impact of forest conversion on RS and its components. The average RH was increased by 35% following the conversion of primary forest to secondary birch(Betula platyphylla) forest,31% to larch (Larix gmelinii) plantation and 19% to pine (Pinus koraiensis) plantation. The average RH revealed a significant positive correlation with soil organic carbon storage in the top soil (R2=0.869,P=0.045), while we did not find significant relationships between the average RH and soil microbial biomass carbon (R2=0.287, P=0.615). When the primary forest was converted to secondary birch forest, the average RR significantly increased (P<0.05), but no significant differences in RR between primary forest and two plantations were detected (P>0.05). The average RR was positively correlated with soil microbial biomass carbon (R2=0.911,P=0.03) and fine and medium root biomass (R2=0.891,P=0.037).Second, RS and its components were measured at the multiple timescales for six typical forest types in northeastern China, including the primary mixed broadleaved-Korean pine forest, spruce-fir valley forest, selective cutting of mixed broadleaved-Korean pine forest, secondary birch forest, Dahurian larch plantation and Korean pine plantation. At diurnal timescales, the RS in the spruce-fir valley forest and the RH in the secondary birch forest and larch plantation revealed a significant positive correlation with soil temperature (P<0.05), while we did not find significant relationships between the RS and its components and soil temperature in the other forest types (P>0.05).Within a day, the amplitude (1.72-2.72 μmol CO2 m-2 s-1) and coefficient of variation (8.0-11.5%) of RH were relatively small in all the forest types except the secondary birch forest and pine plantation, while those (0.56-3.18 μmol CO2 m-2 s-1,36.0-45.7%) of RR in all the forest types were relatively large. The seasonal variations of RS and RH from all forest types followed a bellshaped curve and were similar to that of soil temperature, peaking in July and August. In contrast, the seasonal variation in RR was not completely consistent with that of soil temperature in the six forest types and the coefficient of variation of the RR (38.3-96.5%) was higher than those of Rs (28.0-62.3%) and RH (31.6-55.7%). The Q10 of RR (2.94-4.35) was higher than those of Rs (2.48-2.80) and RH (2.41-2.64), whereas R2 values of exponential model between RR and soil temperature (0.32-0.51) was lower than those of RS (0.62-0.77) and RH (0.71-0.78) in the six forest types. At annual timescales, the coefficient of variation of the RR (12.7-30.5%) was also higher than those of RS (10.2-17.9%) and RH (8.5-14.8%). The RR in the spruce-fir valley forest revealed a significant correlation with air temperature (P<0.05), while we did not find significant relationships between the Rs and its components and climate factors in the other forest types (P>0.05).RS were measured to quantify the variability of RS and identify the controlling factors of the variability of RS at multiple spatial scales (among collars within a plot, among three plots within a specific forest type and among the four forest types within this landscape) for four typical forest types, including the primary mixed broadleaved-Korean pine forest, spruce-fir valley forest, selective cutting of mixed broadleaved-Korean pine forest and Korean pine plantation. The average coefficient of variations of RS within plots (30-52%) were significantly higher than those of RS among plots (20-25%) in each forest type (P<0.05). The variability of RS within a specific forest type was explained by the water-filled pore space, soil organic C content, soil C:N ratio and mean DBH and total basal area of trees within a few metres of the measurement collars. The variability of RS across the four forest types was best explained by the soil C:N ratio (R2=0.63, P=0.001).Finally, RS was measured in the 9 ha plot in 2014 to quantify and visualize the spatial pattern of RS in an old-growth mixed broadleaved-Korean pine forest and identify the roles of biotic and abiotic factors that determine the spatial variation in RS at the ecosystem scale. The variogram analyses revealed 87-91% of the sample variance was explained by autocorrelation over a range of 15 to 23 m during the observation periods. The RS were highly correlated among the measurements made in May, July and September (R=0.60-0.74). The model indicated that the soil water content, bulk density and maximum DBH for trees within 3 m (radius) of the measurement collars explained 46% of the spatial variation in RS seasonally averaged across three observations. RS trended to increase from valley to ridge and with slope degree rising. RS on shady (NE, N) and semi-shady (E, NW) slopes were higher than on sunny (S, SW) and semi-sunny (SE, W) slopes, with the lowest on the flat. |
PDF Full Text Request