doi:10.1016/j.soilbio.2008.11.015
Copyright © 2008 Elsevier Ltd All rights reserved.
Plant carbon inputs and environmental factors strongly affect soil respiration in a subtropical forest of southwestern China
References and further reading may be available for this article. To view references and further reading you must
purchase this article.
Douglas A. Schaefera, 
, 
, Wenting Fenga and Xiaoming Zoua, b
aForest Ecosystem Research Center, Xishuangbanna Tropical Botanical Garden, Chinese Academy of Sciences, 88 Xuefu Road, Kunming, Yunnan 650223, PR China
bInstitute for Tropical Ecosystem Studies, University of Puerto Rico, PO Box 21910, San Juan, PR 00931-1910, USA
Received 14 May 2008;
revised 9 November 2008;
accepted 17 November 2008.
Available online 16 December 2008.
Abstract
Soil respiration is a large component of global carbon fluxes, so it is important to explore how this carbon flux varies with environmental factors and carbon inputs from plants. As part of a long-term study on the chemical and biological effects of aboveground litterfall denial, root trenching and tree-stem girdling, we measured soil respiration for three years in plots where those treatments were applied singly and in combination. Tree-stem girdling terminates the flow of carbohydrates from canopy, but allows the roots to continue water and nutrient uptake. After carbon storage below the stem girdles is depleted, the girdled trees die. Root trenching immediately terminates root exudates as well as water and nutrient uptake. Excluding aboveground litterfall removes soil carbon inputs, but allows normal root functions to continue. We found that removing aboveground litterfall and the humus layer reduced soil respiration by more than the C input from litter, a respiration priming effect. When this treatment was combined with stem girdling, root trenching or those treatments in combination, the change in soil respiration was indistinguishable from the loss of litterfall C inputs. This suggests that litterfall priming occurs only when normal root processes persist. Soil respiration was significantly related to temperature in all treatment combinations, and to soil water content in all treatments except stem girdling alone, and girdling plus trenching. Aboveground litterfall was a significant predictor of soil respiration in control, stem-girdled, trenched and stem-girdled plus trenching treatments. Stem girdling significantly reduced soil respiration as a single factor, but root trenching did not. These results suggest that in addition to temperature, aboveground carbon inputs exert strong controls on forest soil respiration.
Keywords: Carbon cycling; Girdling; Litter removal; Root trenching; Soil respiration; Subtropical forest
Fig. 1. Environmental factors and leaf litterfall measured in the Ailao Mountains forest, Southwest China, from April 2004 to April 2007. A. Soil temperature maximum and minimum among treatments, at 5 cm depth (°C). B. Gravimetric soil moisture (g water g−1 dry soil) for each of the 8 treatments: CCK = unmanipulated, CNL = control, no litter, CNLR = control, no litter and root trenching, CNR = control and root trenching, GCK = tree stem girdling, GNL = girdling and no litter, GNLR = girdling, no litter and root trenching, GNR = girdling and root trenching (legend extends into panel C). C. Average aboveground leaf litterfall (g m−2 d−1) in the CCK and GCK plots for the 30 days prior to each soil respiration measurement. D. Total aboveground leaf litterfall (g m−2) in the CCK and GCK plots during the 60 days prior to each soil respiration measurement.
Fig. 2. Soil respiration (μmol C m−2 s−1) measured before the initiation of treatments in the Ailao Mountains forest, Southwest China, from August 2003 to January 2004.
Fig. 3. Annual soil respiration fluxes (g CO2–C m−2) for each of the 8 treatments measured in the Ailao Mountains forest, Southwest China, from April 2004 to April 2007. Year 1 = April 2004 to March 2005, year 2 = April 2005 to March 2006, year 3 = April 2006 to March 2007 and Average is for those 3 years. Treatments (TRTS): CCK = unmanipulated, CNL = control, no litter, CNLR = control, no litter and root trenching, CNR = control and root trenching, GCK = tree stem girdling, GNL = girdling and no litter, GNLR = girdling, no litter and root trenching, GNR = girdling and root trenching. Bars are displayed in the order of decreasing average fluxes. Different letters within a year indicate significant differences by Mann–Whitney ranking.
Fig. 4. Soil respiration (μmol C m−2 s−1) and standard error bars (N = 4) for each of the 8 treatments measured in the Ailao Mountains forest, Southwest China, from April 2004 to April 2007. A. CCK = unmanipulated, CNL = control, no litter, CNLR = control, no litter and root trenching, CNR = control and root trenching. B. GCK = tree stem girdling, GNL = girdling and no litter, GNLR = girdling, no litter and root trenching, GNR = girdling and root trenching.
Fig. 5. A. Measured and modeled soil respiration (μmol C m−2 s−1) in the Ailao Mountains forest, Southwest China, from April 2004 to April 2007. CNR = control and root trenching; the treatment for which the multiple linear regression model had the highest adjusted R2 (0.769, see text for details of the model). B. As in panel A, GNR = girdling and root trenching; the treatment for which the multiple linear regression model had the lowest adjusted R2 (0.439, see text for details of the model).
Fig. 6. A. Measured and modeled soil respiration (μmol C m−2 s−1) in the Ailao Mountains forest, Southwest China, from April 2004 to March 2007. CNR–GNLR = control and root trenching minus girdling, no litter and root trenching; the treatment difference for which the multiple linear regression model had the highest adjusted R2 (0.794, see text for details of the model). B. As in panel A, GNR–GNLR = girdling and root trenching minus girdling, no litter and root trenching; the treatment for which the multiple linear regression model had the lowest adjusted R2 (0.349, see text for details of the model).
Table 1.
Multiple linear regression model coefficients, statistics, and annual and 3-year average soil respiration fluxes (g CO2–C m−2 y−1) measured in the Ailao Mountains forest, Southwest China, from April 2004 through March 2007. Treatments (TRTS): CCK = unmanipulated, CNL = control, no litter, CNLR = control, no litter and root trenching, CNR = control and root trenching, GCK = tree stem girdling, GNL = girdling and no litter, GNLR = girdling, no litter and root trenching, GNR = girdling and root trenching. Multiple linear regression models are based on the natural logarithm of soil respiration (μmol m−2 s−1) versus soil temperature (°C), soil water content (g water g−1 dry soil), and leaf litterfall in the 60 days prior to measurement (LF60; g m−2). Only those factors with P < 0.05 are reported; other cells are blank. All ANOVA P values were <0.001. LF60 was not tested in the no-litter treatments (–).
Table 2.
Multiple linear regression model coefficients, statistics, and annual and 3-year average soil respiration fluxes (g CO2–C m−2 y−1) for significant treatment differences measured in the Ailao Mountains forest, Southwest China, from April 2004 through March 2007. Treatments (TRTS): CCK = unmanipulated, CNL = control, no litter, CNLR = control, no litter and root trenching, CNR = control and root trenching, GCK = tree stem girdling, GNL = girdling and no litter, GNLR = girdling, no litter and root trenching, GNR = girdling and root trenching. Multiple linear regression models are based on the natural logarithm of soil respiration (μmol m−2 s−1) versus soil temperature (°C), soil water content (g water g−1 dry soil), difference in water content between treatments and leaf litterfall in the 60 days prior to measurement (LF60; g m−2). Only those factors with P < 0.05 are reported except for two non-significant coefficients in bold; other cells are blank. All ANOVA P values were <0.001.
Corresponding author. Tel.: +86 871 5161199; fax: +86 871 5160916.
Abstract
Purchase PDF (581 K)
E-mail Article
Add to my Quick Links
Cited By in Scopus (1)
Purchase PDF (269 K)
