Intercomparison of Three Generations of Infrared Gas Analyzers

IRGAs (infrared gas analyzers) are used in many ecosystems for long-term monitoring of water vapor fluxes in the surface layer of the atmosphere. While these sensors provide accurate measurements, they are also expensive, which prohibits purchase of newer, possibly higher quality instruments; instead, older sensors may be passed on from one completed project to the next one. These sensors may become outdated, resulting in a systematic bias when compared to newer sensors. In our experiment, we had the opportunity to use LICOR’s original LI-7500 (first available in 1999) sensor alongside the LI-7500A (2010) and the LI-7500RS (2016). In a technical note, the manufacturer demonstrated using simple linear regression comparable performance between the LI-7500 and LI-7500A based on latent heat flux for one month over a rye grass field. Optical improvements to the LI-7500RS are claimed to “directly affect the quality of data collected”. With the necessary WPL corrections, fluxes are affected by the quality of the raw water vapor measurements. The purpose of this study is to evaluate how error in absolute humidity propagates into fluxes for the LI-7500, LI-7500A, and LI-7500RS instruments. An experiment was conducted in Bushland, TX on the northern side of a 5-ha field with an irrigated maize canopy of 2.3 m in height. The LI-7500A and LI-7500RS systems were collocated next to a sonic anemometer (CSAT3) and capacitive thin-film polymer hygrometer (HMP 155A), with the LI-7500 on another tower 26 m due south, similarly collocated with a CSAT3. All sensors were at 4.6 m above ground level. Water vapor spectra were calculated, and concentrations were referenced to the HMP 155A. Half-hourly fluxes were computed for three EC systems. Comparisons were made for the period beginning 15 July 2016 and ending 13 September 2016. For absolute humidity, the LI-7500 RS performed best, primarily due to the sensor’s stability at times in which the other two IRGAs had larger variance and possibly spurious values. Errors were generally larger with greater humidity magnitudes. Over the 61 days, maximum water vapor fluxes were highest for the LI-7500, and lowest for the LI-7500RS, averaging greater than 50 W m^-2 higher than for the other two instruments. Water vapor spectra and careful inspection of meteorological conditions will help to better explain the differences between the three IRGAs.