Abstract. Low-cost, medium-precision sensor (LCS) networks have emerged as a promising approach for CO2 monitoring under complex urban emission conditions. However, the field performance of these LCSs faces significant challenges from environmental factors (e.g., temperature, humidity) and long-term drifts caused by sensor degradation (e.g., light source aging). In this study, we performed 30 months of co-located observations using LCS units (named SENSE-IAP) alongside a Picarro reference analyzer to evaluate long-term field performance, which is essential for the correction and validation of mid–low-cost CO2 observation networks. The environmental correction system we developed effectively corrected the impact of daily environmental changes, which reduced the root mean square errors (RMSEs) from 5.9 ± 1.2 to 1.6 ± 0.5 ppm for SENSE-IAP. The corrections remained robust against seasonal environmental changes, maintaining daily RMSE typically within 1–3 ppm throughout the 30 months of observation. Long-term drifts, commonly occurred in LCS, produced biases of up to 27.9 ppm over 2 years. Furthermore, seasonal drift cycle contributed up to 25 ppm RMSE after 6 months of deployment. Although the environmental correction system could not address these errors, linear interpolation effectively calibrated the long-term drift, reducing the 30-month RMSE to 2.4 ± 0.2 ppm. To improve the accuracy of high-density CO2 networks using such LCSs, we recommend maintaining a calibration frequency preferably within 3 months and not exceeding 6 months, with optimal calibration performed during both winter and summer to ensure accuracy within 5 ppm. These findings indicate that SENSE-IAP instruments can operate long-term without requiring a return to the laboratory for re-calibration or frequent field standard gas calibration, thereby substantially reducing time, labor, and financial costs.
Cai et al. (2025) studied this question.