Are fans required for closed chamber N₂O measurements in grazed pastures?

Introduction

Nitrous oxide (N₂O) is a potent, long-lived greenhouse gas. Globally, agriculture is responsible for over half the anthropogenic N₂O emissions (Tian et al. 2020). In New Zealand, where grazed pasture is the major agricultural land use, direct emissions from animal excreta account for about 64% of agricultural N₂O emissions (Ministry for the Environment 2024).

There are many methods for measuring N₂O fluxes from soils including static and dynamic chambers, automated chambers, and eddy covariance methods. Of these, static chambers, a low-cost and simple-to-operate method, has been used in the measurements of trace gas fluxes between the land surface and atmosphere for a century (Lundegårdh 1927). This method involves placing an open-ended chamber onto the soil (sometimes using a pre-installed collar to ensure a good seal between the base of the chamber and the soil surface), then sealing the chamber and collecting gas samples from within the chamber at different points in time.

The low cost of individual chambers means that a larger number of chambers can be deployed enabling a greater spatial coverage and/or larger number of treatments to be investigated. There is a wide range of design options for chambers as described by Clough et al. (2020). Chamber requirements will depend upon the type of system being measured.

Urine patches are a major source of N₂O emissions in New Zealand’s grazed pasture systems. In New Zealand, chambers of 250 mm diameter (area = 0.05 m²) have been used to determine country-specific N₂O emission factors (de Klein et al. 2003). However, naturally deposited cattle urine patch areas can be highly variable in shape and size with mean values in the range of 0.2–0.3 m², 0.3–0.5 m², and 0.3–0.8 m² for 1, 2 and 3 L patches (Jolly et al. 2021). Therefore, when investigating the effectiveness of management practices at field scale it may be desirable to use larger chambers that allow for the natural expansion of a urine patch (O’Neill et al. 2021).

However, accurate flux measurements require the headspace within a chamber to be well-mixed. Using a fan within the chamber is potentially one method to achieve this although there is also a risk that excessive fan speeds could affect the rate of diffusion of gases from the soil. In addition, the use of fans would increase the cost of each chamber and therefore reduce the number of chambers that could be deployed for a given cost. Currently, there are no definitive guidelines on fan use for a given chamber design although mixing is more likely to be an issue as chamber height increases (Clough et al. 2020).

This study aimed to determine whether the presence or absence of a fan affects the N₂O flux measurements from urine applied to pasture using large static chambers (800 mm diameter).

Materials and methods

Field measurements

We analysed results from two experiments using large (800 mm diameter and 250 mm height with a base inserted ~20–40 mm into the soil resulting in a volume of 105–115 L and ground coverage of 0.50 m²) static chambers to measure N₂O emissions. The chamber was not vented and a rubber seal between the top of the chamber and the base was used to seal the chamber head-space. The top was insulated with 20-mm thick polystyrene, with a port for taking gas samples and a port for inserting a thermometer in order to monitor internal temperature. further details reported in (Hoogendoorn et al. 2018). In both experiments, a battery-operated fan was used (Multicomp MC36325 12 V, 80 mm diameter). The air speed was not measured but should be 0.934 m³ per min according to the manufacturers specifications. Fig. 1 shows a picture of the chamber with the lid open.

Fig. 1.

Chamber set up with lid open that was used in experiments.

Both experiments were carried out at two locations in Massey University dairy farms in the Manawatū region of New Zealand. Farm 1 and Farm 4 are situated in temperate climatic region, with mean annual air temperature and mean annual rainfall ranges of 13°C and 903 mm. One was on well-drained soil (Manawatu silt loam, weathered Fluvial Recent) and the other on poorly drained soil (Tokomaru silt loam, an argillic-fragic Perch-gley Pallic) in the New Zealand soil classification system (Hewitt 1992).

The selected areas for the experiments were fenced off for at least 10–12 weeks before the start of the experiments to avoid interference from uneven deposition of fresh dung and urine inputs and N fertiliser application and to reduce spatial variability. The trial layouts for the poorly drained (Adhikari et al. 2024) and well-drained sites consisted of a randomised block design with included 56 plots each split into seven or 14 treatments (a control, urine only, and urine plus five and 12 different nitrification inhibitor treatments, respectively). For each treatment, there were four replicate chambers for the well-drained site (4 × 14 = 56) and eight replicates (8 × 7 = 56) for the poorly drained site. N₂O chamber technology and gas sampling schedules are consistent with the Global Research Alliance N₂O chamber methodology guidelines (Charteris et al. 2020; De Klein et al. 2003). The flux measurements were made between 10:00 hours and 13:00 hours (local time) in both trials with 50 mL gas withdrawn from three time points: 0 (t0); 30 (t30); and 60 (t60) min after the chamber closure with and without the use of battery-operated fans.

On each sampling day, two background air samples were taken. We used these measurements to determine the effect of fan use on mixing the gas within about 105–115 L chamber headspace (assuming an average height of 210–230 mm above the soil surface).

Results

Fig. 2 shows the distribution of N₂O flux measured for the two different experiments using chambers with and without a fan. In both cases, the medians are similar with and without the fans although the emissions are much smaller in the low emission experiment.

Fig. 2.

Distribution of measured N₂O fluxes across all chambers for (a) low emissions and (b) high emissions experiments.

The Kolmogorov–Smirnov test detected no significant differences in the N₂O distributions with and without a fan in the two experiments (P = 0.62 and 0.906 for the low and high emissions experiments, respectively).

Fig. 3 shows the measured fluxes with and without a fan for the two experiments. In the low emission experiment the N₂O fluxes needed to be log transformed to produce normal residuals. As there were some negative fluxes, a small positive constant was added prior to the log transformation and then the differences were calculated. The differences were not significantly different from zero for all but one of the treatments. This difference became non-significant when the Tukey test was applied to correct for multiple comparisons.

Fig. 3.

Measured N₂O emission with and without a fan in chamber for the different treatments in the (a) low emissions and (b) high emissions experiments. Error bar, one s.e.

The high emission experiment did not require transformation and only showed a significant difference from zero for one of the treatments. This treatment also became non-significant under the Tukey test.

Fig. 4 shows the N₂O measurements from each chamber with a fan plotted against the corresponding measurement without a fan for the two experiments. The points lay close to the 1:1 line although there was a lot more scatter in the low emissions experiment. This is expected as under low flux conditions the flux measurement errors will be larger as a proportion of the measured fluxes. There was potentially some underestimation of fluxes without a fan under low flux conditions.

Fig. 4.

N₂O emission measured for each chamber with and without a fan from (a) low emissions and (b) high emissions experiments. The solid lines show linear regressions between with fan and without fan measurements.

Discussion and conclusions

While there could be a slight under-estimation of fluxes measured without fans under low emitting conditions, our results showed negligible effect of fans on the measured N₂O emissions. In contrast, in a laboratory experiment, Christiansen et al. (2011) found that manual sampling with no headspace mixing resulted in an underprediction of CH₄ fluxes. However, these chambers were taller (0.41–1.82 m) with a smaller base area (0.196 m²) than our studies. The rate of manual sampling was also much faster (every 5 or 10 min) compared to this study where samples were only taken at 30-min intervals. The shorter time interval may have meant there was not enough time for the headspace gas to recover between sampling events without forced mixing.

This suggests the effect of using fans depends on many factors such as the chamber dimensions, flux rates, and the rate of gas sampling. Other authors have suggested that the use of fans should aim to emulate ambient wind conditions (e.g. Rochette 2011; Pavelka et al. 2018). Our results indicate that fans may not be necessary for N₂O flux measurements from pasture soils over the range of flux rates considered, even with quite a large chamber headspace. However, these results might not hold for systems with taller chambers containing more plant biomass.