Spatial heterogeneity of soil respiration in a Larix gmelinii forest and the response to prescribed fire in the Greater Xing′an Mountains, China

Abstract

This study was conducted in a fire-prone region in the Greater Xing'an Mountains, the primary forested area of northeastern China. We measured soil respiration and the affecting soil factors, i.e., soil microbial biomass and soil moisture, within an experimental plot of Larix gmelinii Rupr. A low-intensity, prescribed fire was applied as the treatment. Traditional descriptive statistics and geostatistics were used to analyze the spatial heterogeneity of soil respiration and the response of respiration to fire disturbance. Coefficients of variation (CVs) for pre-fire and post-fire soil respiration were 23.4 and 32.0 %, respectively. CVs for post-fire soil respiration increased significantly, with a moderate variation of all CVs. Soil respiration pre-fire was significantly correlated with soil microbial biomass carbon, biomass nitrogen, and soil moisture (W); post-fire soil respiration was not correlated with these factors. From the geostatistical analyses, the C 0 ? C (sill) for post-fire soil respiration increased significantly, indicating that the post-fire spatial heterogeneity of soil respiration increased significantly. The nugget effect (n c ) of soil respiration and the affecting factors pre-fire and post-fire disturbance were in the range of 12.5-50 %, with strong spatial autocorrelation. Fire disturbance changed the components of spatial heterogeneity, and the proportion of functional heterogeneity increased significantly post-fire. The ranges (a) for pre-fire and post-fire soil respiration were 81.0 and 68.2 m, respectively. The homogeneity of the distribution of post-fire soil respiration decreased and the spatial heterogeneity increased, thus the range for postfire soil respiration decreased significantly. The fractal dimension (D) for soil respiration increased post-fire, the spatial heterogeneity of soil respiration affected by random components increased, indicating that the change in spatial heterogeneity of post-fire soil respiration should be considered within the scale of the forest stand. Following Kriging interpolation, the increase in the patchiness of post-fire soil respiration was illustrated using a contour map. Based on these preliminary results, the change in the spatial heterogeneity of post-fire soil respiration was likely caused by changes in the distribution of soil moisture and microbial activity within the experimental plot at the scale of the forest stand.

Short- to medium-term effects of crown and surface fires on soil respiration in a Canadian boreal forest

“…The R S measurements conducted immediately after the non-stand-replacing surface fire in 2018 ranged from 9.7 to 43 g CO 2 m -2 d -1 . This is substantially higher than the mean CO 2 efflux (3.4 g CO 2 m -2 d -1 ) registered immediately after a prescribed low-intensity surface fire in a larch forest in China (Hu et al, 2016). Conversely, our measurements of the R S (6.5 to 20.8 g CO 2 m -2 d -1 ) 1 -3 years after the stand-replacing surface fires were close to mean CO 2 effluxes measured 4 (15.7 g CO 2 m -2 d -1 ) and 5 years (4.1 g CO 2 m -2 d -1 ) recorded after stand-replacing surface fires in larch forests in eastern Siberia (Sawamoto et al, 2000;Takakai et al, 2010).…”

Seasonal variations in temperature sensitivity of soil respiration in a larch forest in the Northern Daxing’an Mountains in Northeast China

“…Q 10 values could also be indirectly affected by the frequency and intervals of soil respiration measurements (Chen et al 2020). In a measurement process without a specific period (freeze thaw period, rainy season) and long measurement intervals, the calculated Q 10 values tend to be lower (Peichl et al 2014;Hu et al 2016). At the same time, the observed Q 10 values are also affected by soil depth (Li et al 2020).…”

Leveraging explainable machine learning models to assess forest health: A case study in Hainan, China

“…In this regard, it is imperative to develop models that are capable of accurately assessing the health of forests in the future. Previous studies have used various methods to assess forest health, including principal component analysis (PCA), the health distance method (HDM), and comprehensive evaluation methods (Kayet et al, 2019 ; Tao et al, 2019 ; Xue et al, 2013 ; Zhu et al, 2013 ). However, all of these methods suffer from low classification accuracy, so some researchers have tried to solve this problem with machine learning algorithms (Ki, 2019 ).…”