The eddy covariance method provides valuable information to better understand temporal variability and estimates an annual CH4 emission average that can be compared to inventories. Only a select number of studies have conducted in situ field measurements of CH4 from California dairy manure lagoons. The magnitude and temporal patterns of CH4 emissions from manure lagoons often vary depending on the method used to estimate emissions. There is also an important role of seasonality of CH4 emissions that might confound comparison of atmosphere-based estimates with inventory. For example, Arndt et al. showed that summer CH4 emissions were comparable to inventory estimates, but not during winter measurements. In addition, emissions from manure liquid storage were 3 to 6 times higher during the summer measurements than during the winter measurements using three different techniques . In a recent study, statewide emission factors were comparable to ground-level measurements during the summer and fall seasons, but airborne measurements were 8% higher than the statewide inventories . Methane emissions from dairy manure lagoons may also differ by as much as a factor of two using different dispersion models . Other important gaseous emissions are also co-emitted with CH4 at dairy farms, but have different spatial patterns because they are coming from different sources . Additional observations at the seasonal and diel scales are needed to address uncertainties in CH4 emissions from dairy manure lagoons in California. In this study, we investigate seasonal and diurnal CH4 fluxes from manure lagoons at a dairy farm in Southern California using the eddy covariance technique. We pair our CH4 fluxes with micrometeorological measurements, air racking including wind speed, surface pond temperature, air temperature, among other parameters.
We then discuss the impact of lagoon agitation events, such as precipitation and manure management practices, on CH4 fluxes. Finally, we compare our CH4 flux estimates using the eddy covariance technique with other methods deployed at the same location. We hypothesized that manure lagoon CH4 emissions would follow seasonal patterns, with higher fluxes in spring and summer when manure substrate availability and temperature are higher. We also surmised that higher wind speeds would increase CH4 fluxes through increased turbulence and mixing of the lagoon surface. Finally, we hypothesized that manure management practices would have a measurable impact on measured CH4 emissions.Our study site is a manure storage lagoon on a typical dairy in southern California, located near 33.8º, -117.0º . The site has a semi-arid climate, with a mean annual temperature of 19⁰C and mean annual precipitation of 0.5 ± 2.6 mm that mostly falls between November and March. It is an open dry lot dairy—meaning that milk cows are housed in open corrals with dirt surfaces, and manure deposited in feed lanes is primarily scraped off the lot rather than flushed with water. The manure that is scraped from the corrals is stored as dry manure piles south of the dry lot. Water is used to flush out manure deposited in the milking parlor into manure ponds via the subsurface and above ground channels . Corral runoff flows to the channels via drainage pits,with four weeping walls present to retain solids. Approximately 227,100 L of storm water runoff from corrals and feed lanes , milk parlor wash down water, and wash pen water enters the manure pond system daily. Manure ponds receive about 38,000 L of fresh dairy flush manure daily. From December 2016 to June 2018, 56,775 L per day of green waste digestate was also introduced to the manure lagoons for testing their Ag Waste Solutions system that converts cow manure into bio-fuel, primarily diesel fuel, and bio-char . Occasionally, solids are removed from the above ground channels and stored as dry manure storage piles . The dairy farm’s population consist exclusively of Holstein cows. Demographics are relatively stable between seasons since it is a closed herd—births are on site and cows only leave once they retire or pass away.
There are approximately 1066 milking cows, 200 dry cows, 685 heifers, and 370 calves. The dairy manure flush system only receives input from the milking cows and calves. The total annual manure produced from dry corral production is 6300 tons.The manure pond system consists of five manure ponds , wherein the liquid manure navigates from manure pond 1 to manure pond 5 via gravity, decreasing the content of suspended volatile solids through anaerobic decomposition and settling as it navigates from one manure pond to the next. Throughout the study period , the surface of manure pond 1 underwent a drastic change in vegetation and surface variation . To quantify the percentage change in crust/vegetation,we calculated the change in vegetation/crust area using Google Earth satellite imagery between 2019 and 2021. There was a 147% increase in area covered by the crust layer and vegetation on manure pond 1 from June 2019 to June 2021. Peak vegetation growth occurred during the summer months , followed by a dry period. We define the pre-sedimentation stage occurring from June 2019 to May 2020 and the postsedimentation stage occurring from June 2020 and June 2021 when a substantial crust and sediment layer formed on the surface of manure pond 1. A common practice is to dredge dairy manure ponds periodically. However, the Southern California dairy farm has not dredged their manure ponds since it was constructed in 2006, thus solids also accumulated throughout this study period. The solids in the channel leading to the manure pond system were dredged in March 2020 following rain events and December 1, 2020 . Typically, the channels are dredged twice a year.We installed an eddy covariance flux tower at a height of 4 m on the southeastern edge of Lagoon 1 . The eddy covariance flux tower consisted of an open-path CH4 analyzer , integrated CO2 and H2O Open-Path Gas Analyzer and 3-D Sonic Anemometer . The analyzers measured at a rate of 10 Hz. They were calibrated before and after the field measurements using zero air and custom gas mixtures that were tied to the scale set by the NOAA Global Monitoring Division by measurement against NOAA certified tanks. We also measured air temperature and relative humidity , the surface temperature of the pond with an infrared radiometer , and precipitation with a rain gauge.
The data were recorded using a CR3000 datalogger. Instruments were powered using three solar panels, seven deepcycle. Dust was removed using an automatic cleaning system.The footprint of an eddy covariance flux measurements represents the upwind area that contributes to the fluxes at the location of measurements. The extent of the footprint depends on the micrometeorological conditions such as stability of the boundary layer and wind speed. A flux footprint model by Kljun et al. was used to estimate the footprint of the eddy covariance flux measurements. The algorithm uses the following inputs to calculate the footprint: mean wind speed, wind direction, weed dryer standard deviation of the horizontal wind speed, friction velocity, planetary boundary layer height, and Obukhov length. Figure 4.3 shows the upwind area that contributes to the flux observations with friction velocity greater than 0.1 m s-1 , wind direction between 270⁰ and 340⁰, and wind speed greater than 0.2 m s-1 . The distance of footprint contributions were calculated for each half-hour flux using the EddyPro software. The extent of the footprint captures manure pond 1, manure pond 2, and a portion of manure pond 3. As shown in Figure 4.3, 70% of the footprint primarily covers less than 50% of the area of manure pond 1.On August 28, 2019, we sampled the manure lagoon complex for various biophysical parameters using a boat at three different locations and depths. We sampled at three locations shown in Figure 4.4 and Figure 4.10. L1 and L2 were sampled at 0 and 0.3 m and L3 was sampled at surface level, 0.3, and 0.8 m. L1 and L2 were only sampled at the surface level and 0.3 depth since the high volatile content limited the instrumentation’s reach. We measured pH and temperature with an Oakton PCTS 50, PCSTestr 35 or pHTestr 30. Oxidation-reduction potential was measured with anOakton ORP Testr 10 that was calibrated with Zobell’s solution from VWR Scientific in the lab 24 hours prior to field work. Electrical conductivity was measured on each liquid sample in the laboratory using an Oakton Con 100 series meter and conductivity probe. The probe was calibrated according to manufacturer’s recommendations with 1413 uS standard solution from Fisher Scientific. Samples were removed from the 4 °C cold room and each was inverted gently 2-3 times to mix contents just prior to measurement. The probe was calibrated after every 10-15 readings to reduce drift. Total solids concentration , which is the solid concentration of biomass, was determined by weighing and drying 15-25 ml aliquots of each sample in triplicate in a 120 °C oven for 4- 16 hours, weighing the residual, then dividing by the wet weight. Aliquots were made using the shake and pour method . Fixed solids concentration , which is the inorganic fraction of total solids, was determined by further combustion of the dried samples in a muffle oven at 540 °C for 4 hours, weighing the residual, then dividing by the dry weight . Volatile solids concentration , which is the organic fraction of total solids, is the difference between TS and FS divided by wet weight. On August 14, 2018, stationary measurements of CH4 mole fractions downwind of manure pond 1 were collected with a cavity-ring down spectrometer . Dispersion models were then used to estimate CH4 emissions and showed that CH4 emissions were heterogenous, with higher CH4 emissionsnear the manure stream inlet . In a pilot study on August 27, 2019, CH4 emissions were estimated using an auto-ventilated floating chamber connected to a CRDS .
Figure 4.5 shows the timeline of measurements conducted at manure pond 1.During the study period, observed air temperatures were on average 19 ⁰C, with the highest temperatures measured during the summer . Sensible heat flux was on average 41 W m-2 . Mean surface pond temperatures were comparable to mean air temperatures with 20 ⁰C. Friction velocity was on average 0.2 ± 0.1 ms-1 . Lastly, incoming shortwave radiation near the manure ponds was 75±71 Wm-2 , on average . In our study site, precipitation events were highest during the winter and spring seasons. The highest precipitation events occurred during March and April in the year 2020. Daily CH4 fluxes were also highest during this time . Surface and pond temperatures were on average highest during the summer months of August and September. Similarly, incoming shortwave radiation was strongest during the summer months of August and September in the year 2020. There were no overall seasonal patterns observed for friction velocity and wind speed.At the diurnal scale, the micrometeorological factors that had the strongest correlations with CH4 fluxes were air and surface pond temperature, wind speed, and friction velocity based on linear regression models. The micrometeorological factors that had the strongest effects differed between the pre-sedimentation stage and post-sedimentation stage . There was a strong diurnal relationship between CH4 fluxes and surface pond temperature fluxes, especially during the pre-sedimentation stage of manure pond 1 . However, the diurnal connection between CH4 fluxes and pond temperature weakens postsedimentation . Methane fluxes and latent heat fluxes follow a similar diurnal pattern , with peaks during the early afternoon, when pond and air temperatures were also the highest . Wind speed also had a significant effect on diurnal CH4 fluxes during both pre-sedimentation and post-sedimentation conditions. Friction velocity had a stronger influence on CH4 fluxes during the post-sedimentation phase than during the pre-sedimentation phase of manure pond 1. During our study period, CH4 fluxes from manure pond 1 decreased from 2019 to 2021, with the highest CH4 fluxes observed during the spring period in 2020 . Spring CH4 fluxes decreased on average by 70% from 2020 to 2021 and summer CH4 fluxes decreased on average by 57% from 2019 to 2021. Monthly CO2 fluxes increased during the spring season and then decreased during the summer months, when there was vegetation growth in manure pond 1, driving photosynthesis and carbon uptake. Methane fluxes and CO2 fluxes followed a similar seasonal pattern . In contrast to diurnal CH4 fluxes, seasonal CH4 fluxes were not significantly correlated with seasonal latent heat flux . Monthly latent heat fluxes increased during the summer, whereas monthly CH4 fluxes decreased during the summer.