Assessing the relative importance of soil moisture and temperature on the nighttime CO2 flux: two contrasted ecosystems study cases in West Africa

In the context of global warming and rapid increase of population particularly in West Africa where forest ecosystems are threatened by land use conversion, understanding the biophysical variables influencing the ecosystem respiration (Reco) becomes vital for predicting the carbon balance in response to climate change. Using the Eddy Covariance method, and without an automatic chamber system, a key step to determine Reco is to use nighttime Net Ecosystem Exchange (NEE) data to establish a functional relationship using Reco main driver’s as an input.  However, to ensure that the input variable(s) used in this relationship is the most relevant governing this process, especially in tropical regions where both data and studies are scarce, it remains therefore an important prerequisite to examine the relative importance of potential drivers. This prior analysis is especially needed because: (1) the main driver may differ according to the climate and locations; (2) collinearity between some potential drivers could be a complex issue in identifying the main one; (3) the scale of influence of these drivers on the nighttime CO2 emission at some sites is still unknown. In our study, we therefore investigated the relative importance and scale of influence of soil moisture (Hsoil) and temperature (Tsoil) at different depths on the nighttime NEE. Since variations of the net CO2 flux exchanged can differ from one site to another, in this study we used data acquired from 2008 to 2017 above two contrasted ecosystems: a mixed culture (Nalohou, lat. 9.74^°N, long. 1.60^°E) and an open clear forest (Bellefoungou, lat. 9.79°N, long.1.72°E) located in Sudanian climate, Northern Benin. Both sites belong to the AMMA-CATCH (African Monsoon Multidisciplinary Analysis-Coupling of the Tropical Atmosphere and Hydrological Cycle) observatory. Two methods have been then deployed: the first one which is the Mutual Information, was used to identify the relative importance of Hsoil and Tsoil in controlling the nighttime NEE; the second one, the wavelet transform allows determining the scales of influence of these variables on nighttime NEE. We found that periodicity of nighttime NEE response to the two variables differs according to the season.  Soil moisture appears as the most important variable in the nighttime NEE variation whatever ecosystems and seasons analyzed. During wet seasons, nighttime NEE response to soil moisture exhibits a periodicity lower than 64 nights with synchronization every 07 nights, while this synchronization can extend from 12 to 14 nights during the dry season. This fast response of nighttime CO₂ emissions to soil moisture during the wet season results from a significant increase in precipitation. During the dry seasons, without precipitation, soil moisture decreases, and this reduces the water available to plant growth and microorganism activity, thus reducing the amount of CO2 emitted. Moreover, sporadic rainfall events rewet the soil, leading to spontaneous CO2 emissions. Soil temperature is secondary in importance, but its impact can vary; it is less relevant during the dry season, or more redundant in the wet season, for both sites. It is also out of phase with nighttime CO2 emissions regardless of the season.