Portable infrared CO
2 sensors, such as GMT222 (VAISALA) and LI-820 (LI-COR), are often used for soil CO
2 efflux measurements; however, in general, these inexpensive small sensors are relatively inaccurate and slow to respond. The output signals are delayed as a result of diffusion processes in the sample cell and internal averaging calculations to ensure stable data output. Accurate estimations of CO
2 efflux require calculating the actual increase in CO
2 concentration in a chamber without these composite delays.
Laboratory experiments were conducted to determine the response characteristics of several sensors under two different conditions: diffusion and flow-through. The data obtained in the experiments indicated that attachment of a dust filter substantially delayed the sensor response. Furthermore, the sensor response time under the diffusion condition was longer than under the flow-through condition. When the soil CO
2 efflux was calculated using a nonlinear regression method, a long response time led to large CO
2 efflux estimation errors; for example, a 10% underestimation occurred when CO
2 concentrations were measured using GMM222C with a sampling interval of 300 seconds. We developed a backward estimation method for recovering the actual increase in CO
2 concentration from the delayed sensor outputs. Correction of the sensor response with the backward estimation provides an effective solution for achieving an accurate estimation of the soil CO
2 efflux using inexpensive sensors.
抄録全体を表示