Continuous, long-term monitoring of soil CO2 concentration and CO2 flux using a novel, low-cost CO2 sensor system

Soil respiration (F_s) datasets often exhibit low temporal-spatial resolution and spatial bias, particularly lacking observations in arid/semi-arid regions. This limitation significantly constrains our understanding of the mechanisms governing soil carbon dynamics and hinders the correct estimation of CO₂ emissions at regional to global scales. Challenges in Fs estimation arise mainly from logistical constraints in manual data collection and the high costs of commercial measuring devices. To address this, we developed a low-cost, open-source, autonomous soil CO₂ sensor system. The system design emphasized easy adoption and customization for non-engineer end-users, enabling the collection of high-frequency, long-term soil CO₂ concentration data, and consequently, Fs estimates. A system including six low-cost CO₂ sensors distributed at two soil depths (5 and 10cm) was deployed in the Negev Desert since May 2023. Fs estimates were determined from CO₂ concentration gradient using Fick's law (F_G) and cross-validated with Fs measured by automated chambers (F_C). We found a good agreement between F_G and F_C both in the short term (i.e., sub-daily fluctuation) and long term (i.e., annual net CO₂ emission). Our data also revealed daily and seasonal Fs patterns correlating with environmental factors like temperature and precipitation. The results demonstrate that our system, despite costing less than 10% of automated chamber systems, offers equivalent accuracy in Fs estimates, higher temporal resolution, and potential for enhanced spatial resolution if widely adopted.