Comparing Results of High and Low Frequency Measurements of Greenhouse Gas Fluxes

The emission of greenhouse gasses from soils is caused by dynamic factors such as temperature and moisture. These two factors are also influenced by a multitude of external factors such as diel cycles, land management, wetting/drying events and seasonality; which all lead to a high level of variability in greenhouse gasses emitted (Savage et al., 2014). The result is that greenhouse gas fluxes from soils can be highly variable both temporally and spatially. Therefore, it greatly enhances a scientific study to incorporate equipment best suited for the measurements being taken, and to account for variability that may exist. When seeking to quantify soil gas fluxes, data collection methods that account for these variabilities is key to accurate results; as results are only as accurate has the measurement devices. Continuous measurement devices are the best way to account for spatial variability, and in Savage et al., 2014, they compare measurements of carbon dioxide, methane, and nitrous oxide made with both a manual sampling and continuous sampling methods. This was done using a static chamber recording data three times per week with a continuous measurement chamber that uses a quantum cascade laser (QCLAS), which measures both frequently (once per hour) and is capable of detecting gasses at concentrations 10 times lower than the manual system. The results from this study speaks volumes to the value of using continuous measurement for achieving the best results in gas flux studies.

In the study, Savage et al. sampled the fluxes of three greenhouse gasses mentioned above in a forested wetland at the Howland Forest Research Site, located near Bangor, Maine and an alfalfa field near Mandan, North Dakota. At each site, automated chambers measured from September to November 2011 in the forested wetland and for a full crop cycle beginning in March 2012 at the alfalfa site. Other data that was measured including rainfall and temperature. The automated chambers recorded data once per hour, which involved a two-minute flush of the tubing followed by the chamber lowering and sealing itself for eight minutes, then measuring the gas concentrations before opening again. The figure below demonstrates the sampling process, note that chambers are not to scale.

Screenshot (6)

The manual sampling technique consisted of circular, vented chambers that were placed in the ground to a depth of 10 cm.  They were closed for 6 minutes before sampling, taking 3 samples 3 times per week. Samples were analyzed immediately using gas analyzers.


Automated Chambers in Howland Forest

The minimum detectable fluxes (MDF) were calculated using the method by Parkin et al., 2012, which found the minimum detectable fluxes for the manual chamber to be ±0.70ug N2O-N m-2 h-1, ±3.32ug CH4-C m-2 h-1. In comparison, the MDF for the manual chamber was ±0.05ug N2O-N m-2 h-1, and ±0.18ug CH4-C m-2 h-1. This precise measurement tool allowed for detection of small and sporadic production of N20 at the forested site, where consumption of N20 dominated the soil atmosphere exchange. At the alfalfa site, consistent N20 release from soils were measured. The wetland site measured a small production or consumption of CH4 depending on other factors, particularly fluxes neared zero after rainfall. There was little correlation in the variations of CH4 flux at the alfalfa site. CO2 flux at both sites showed more variation in concentrations observed with the automated chambers than the manual chambers.


Automated Chamber in Howland Forest

Savage et al. showed that there are several concerns when using manual chambers where measuring is done infrequently, as variability of greenhouse gas flux can be high and often has a sporadic or cyclical nature. Without measuring frequently to account for these factors, results will vary from reality and in this study measurements tended to be higher in the manual chambers. Having confidence that chambers are accounting for smaller variations is important when using data to model changes in flux over regional or global scales, where uncertainty will be magnified.