Soil respiration dynamics: understanding and predicting soil CO₂ flux of dry deciduous forest ecosystem in India using Q₁₀ models and random forest
摘要
Soil CO₂ flux, the exchange of carbon dioxide between the soil and atmosphere driven by processes like microbial decomposition and root respiration. This study investigates the diurnal and seasonal patterns of soil CO₂ flux and their relationship with environmental factors particularly soil temperature and soil moisture. There were visible patterns mostly pronounced between March and June, peaking shortly after noon, driven by rising soil temperatures. During the non-rainy season, the average daily soil respiration flux was 0.143 g m⁻² day⁻¹, with a variation of ± 0.03 g m⁻² day⁻¹. In contrast, the monsoon season exhibited higher variability, averaging 0.203 ± 0.14 g m⁻² day⁻¹, likely due to increased microbial activity from rainfall-enhanced decomposition. Monthly flux analysis revealed minimal variation during colder months (± 0.017 g m⁻² day⁻¹) and greater variability in warmer months, exceeding ± 0.0259 g m⁻² day⁻¹. Soil temperature was identified as the dominant driver from January to June, when rainfall is limited, while soil moisture exerted greater influence following the onset of the monsoon (June–December). Two variations of the Q₁₀ model – incorporating soil temperature and moisture as dependent variables – along with Random Forest models were evaluated for predicting soil respiration. The Random Forest model with eight input variables outperformed the others, with R² = 0.89, RMSE = 0.41 µmol m⁻² s⁻¹, MAE = 0.24 µmol m⁻² s⁻¹. The total annual soil respiration was estimated at 794 ± 114 gC m-2 yr− 1. These findings highlight the complex interplay of environmental factors in regulating soil respiration and emphasize the importance of considering more than temperature and moisture dynamics in soil carbon flux modelling.