Infrared Gas Analyzers vs. Gas Chromatography in Coastal Ecosystems

Out with the old, in with the new?

A recent study by Brannon and colleagues (2016) out of the Moseman-Valtierra Lab aimed to compare ye olde gas chromatograph (Shimadzu GC-2014) with the newer technologies of a cavity ring down spectroscopy (CRDS) gas analyzer (Picarro, Model G2508) and an off-axis integrated cavity output spectroscopy (OA-ICOS) analyzer (Los Gatos Research, Model N2O/CO) in a coastal marsh.


Salt marsh field site set up. Photo: Picarro Inc.

Coastal ecosystems have considerable carbon storage capabilities, but unlike peatlands, their terrestrial counterparts, coastal wetlands have minimal methane (CH4) and (N2O) emissions. That being said, these greenhouse gas (GHG) fluxes can have large spatial and temporal variability due natural (tidal and diel cycles, temperature changes, flood gradients) and anthropogenic events (ie: nutrient loading), which can promote emissions of both CH4 and N2O in quantities sufficient enough to offset portions of CO2 uptake. This highlights the importance for having in situ, continuous measurements for these gas species in order to get a better picture of the GHG dynamics occurring in these coastal ecosystems.

Brannon et al. (2016) point out that the few studies that have compared CRDS or OA-ICOS infrared analyzers with GC have been agriculture-based. In order to fill in this gap for coastal ecosystem GHG flux dynamics, they set out to study three specific objectives:

1) Minimum Detection Limits for all (Picarro G2508, LGR analyzer, and Shimadzu GC-2014)

Using linear rates of change in gas concentrations (as per Martin and Moseman-Valtierra 2015), minimum flux detection limits were calculated for all analyzers. The Picarro G2508 had a lower analytical detection limit for all three gases (CO2, CH4, N2O), at 1-3 orders of magnitude lower than the Shimadzu GC-2014, as well as greater precision (see table below).


Calculated minimum detectable flux results from field experiments (Brannon et al. 2016).

Sampling from salt marsh mesocosms, the Picarro G2508 was able to consistently detect small CO2 and CH4 fluxes (2 µmol m-2 s-1, and 1 µmol m-2 h-1, respectively), which were below that of the limit of detection of the GC. This finding is corroborated by Christiansen et al. (2015) while sampling soils from forest, wetland and agricultural soils.

The authors state that the majority of CH4 and CO2 fluxes were below the limit of detection for the Shimadzu GC-2014, however it is important to note that longer chamber closure times decreases the discrepancy between the detection limits of the Picarro and GC.

2) Picarro G2508 vs. Shimadzu GC-2014: CO2, CH4, and N2O mesocosm flux comparison

This study used 4-5 minute long chamber closure times, however when using 30 minute closure times the GC’s detection rates increased to 70 µmol N2O per m-2 h-1, which were comparable to those of the Picarro for that time period. It is important to note that longer sampling hours are introduced with longer chamber closure times, such as alterations to gas diffusion gradients, and increased temperatures.

3) Picarro G2508 vs. LGR analyzer: N2O fluxes in mesocosm and field experiments

When comparing the two infrared (IR) gas analyzers’ N2O flux measurements, they were generally similar in both mesocosm and field experiments. The Picarro fluxes were only slightly larger than that of the LGR, at 9-13% difference, in some mesocosms and field plots where nitrogen additions were low, and fluxes were relatively low (3-132 µmol m-2 h-1). Possible explanations offered by Brannon et al. (2015) for this are having a low sample size, as no difference was found between the two analyzers for N2O fluxes of high nitrogen field plots, and also possibly due to different ranges used by the two analyzers; near-IR for Picarro, and mid-IR for LGR.


Photo: Moseman-Valtierra Lab

While the IR gas analyzers are appealing because of their low detection limits, simultaneous gas species measurements, and real-time flux measurements, they do come with some cons particularly for working in coastal ecosystems; they are particularly sensitive to water and moisture condensation. Even small amounts of condensation on the mirrors can lead to extensive repairs to your system. One bonus with the Picarro G2508 is that it monitors moisture levels and alerts the user if a set threshold is exceeded. The G2508 also uses two hydrophobic membrane filters, which trap water droplets in the inlet system before it reaches the optical cavity. Even without these two handy features, Brannon et al. (2016) suggest switching the inlet and outlet tubing if moisture begins to rise in the system, and recommend designing simple moisture traps if working in more humid conditions to avoid any issues.

Real-time measurements and dramatically higher flux detection limits offered by the IR gas analyzers give the coastal scientist a large advantage over using grab sample based GC methods. The ability to detect rapid changes in gas concentrations and fluxes offers the ability to detect experimental errors and allows the user to repeat samples if necessary.

Feel free to read the publication for yourself for more detailed information on the methods used in the different comparisons.


Brannon, E. Q., Moseman-Valtierra, S. M., Rella, C. W., Martin, R. M., Chen, X., and J. Tang (2016) Evaluation of laser-based spectrometers for greenhouse gas flux measurements in coastal marshes, Limnology and Oceanography: Methods 00: 1-11.

Christiansen, J. R., Outhwaite, J., and S. M. Smukler (2015) Comparison of CO2, CH4 and N2O soil-atmosphere exchange measured in static chambers with cavity ring-down spectroscopy and gas chromatography. Agriv. For. Meteorol. 211-212: 48-57.

Martin, M., and S. Moseman-Valtierra (2015) Greenhouse gas fluxes vary between phragmities Australis and native vegetation zones in coastal wetlands along a salinity gradient, Wetlands.