Using an advanced robotic chamber system to detect spatio-temporal short-term responses in measured CO2 exchange to soil manipulation and N fertilization

Improved agricultural practices sequestering additional atmospheric C within the soil are considered as one of the potential solution for mitigating global climate change. However, agricultural used landscapes are complex and their capacity to sequester additional atmospheric C differs substantially in time and space. Hence, accurate and precise information on the complex spatio-temporal CO₂ flux pattern is needed to evaluate the effects/benefits of new agricultural practices aiming towards increasing soil organic carbon.

To date, different approaches are used to measure and quantify CO₂ flux dynamics of agricultural landscapes, such as e.g. eddy covariance, as well as manual and automatic chamber systems. However, all these methods fail to some extend in either accounting for small scale spatial heterogeneity (e.g., eddy covariance and automatic chambers) or short-term temporal variability (e.g., manual chambers). Although, automatic chambers are in principle capable to detect small-scale spatial differences of CO₂ flux dynamics in a sufficient temporal resolution, these systems are usually limited to only a few spatial repetitions which is not sufficient to represent small scale soil heterogeneity such as present within the widespread hummocky ground moraine landscape of NE-Germany.

To overcome these challenges, we developed a novel robotic chamber system. This system was used to detect small-scale spatial heterogeneity and short-term temporal variability of CO₂ flux dynamics in a full factorial experimental setup for a range of three different soil types, two N fertilization forms (2; mineral vs. organic) and two soil manipulation status, representing two different tillage practices. Here, we present measured CO₂ flux dynamics and cumulative emissions for the 3 repetitions of the 12 randomized treatments (36 subplots) directly following soil manipulation and N fertilization during summer 2020. Our results show distinct differences between the three measured soil types as well as a clear response of all three soil types to conducted soil manipulation, yielding in significantly lower ecosystem respiration (R_eco) and net ecosystem exchange (NEE) for manipulated vs. non-manipulated subplots. No clear difference, however, was obtained in case of N fertilization.