UC Berkeley

Landfills are known to be a significant contributor to atmospheric methane, yet emissions are difficult to quantify because they are heterogeneous over a large area (up to 1 km^2) as well as unsteady in time. Many different measurement methods have been developed, each with limitations and errors due to various factors. The most important difference among different measurement methods is the size of the measurement footprint. Flux chambers have the smallest footprint, typically 1 m^2, radial plume mapping mass balance and eddy covariance can have footprints between 100 and 10,000 m^2, and aircraft-based mass balance and the tracer dilution method can measure whole landfill emissions. Whole landfill measurement techniques are considered the best because they account for spatial heterogeneity, and the tracer dilution method (TDM) in particular has gained popularity because it is relatively noninvasive and cost-effective.

The TDM works by comparing ratios of methane and tracer gas plumes downwind of the landfill. A tracer gas such as acetylene is emitted from the landfill at a known, steady rate. Downwind plume transects are collected using a gas analyzer on a moving vehicle to obtain both methane and tracer gas concentrations. The idea behind the method is that at the transect measurement location, the gas plumes are well mixed enough that the ratio between the methane and tracer gas concentrations is approximately equal to the ratio between the methane and tracer gas emissions rates. Methane emissions are calculated by equating the ratio of concentrations to the ratio of emissions rates and solving for methane emissions. Field studies of the TDM with controlled methane releases over a flat field have quantified TDM-related measurement errors, but these errors are impossible to assess in real landfill measurements because the true landfill emissions are unknown.

Numerical modeling of the TDM is an advantageous way to study the errors in TDM- measured emissions as well as how the error changes under a variety of different conditions. With a numerical model, the true methane emissions are prescribed, so they can be compared to the TDM-measured emissions to evaluate the method’s accuracy. The TDM is examined in this dissertation using numerical simulations with the Weather Research and Forecasting model (WRF). WRF is a mesoscale numerical weather prediction model with large-eddy simulation (LES) capabilities, allowing for high-resolution simulations with resolved large- scale turbulent motions. To examine the TDM, two real landfills in the U.S. are selected, with high resolution topography data and real atmospheric data informing the initial and boundary conditions of the model. The simulations are run with a nested grid configuration, starting from 2.25 km resolution and nesting to 150 m resolution and then 30 m horizontal grid resolution over the landfill area, with the lowest vertical level ∼15 m. To simulate the TDM, three components are needed: tracer emissions with a specified configuration and emissions rate, landfill emissions specified at every grid point on the surface of the landfill, and simulated transect measurements with a specified transect path and transect collection speed. In this dissertation, tracer emissions are prescribed as steady values at grid points on the surface of the landfill constituting various tracer configurations to be examined, and methane emissions are either prescribed as steady values or calculated using prescribed soil methane concentrations and a surface flux parameterization.

To our knowledge, the work presented in this dissertation constitutes the first time WRF simulations have been used to examine the TDM. In the first part of this dissertation, steady landfill methane emissions are prescribed to study the effects of various aspects of the TDM setup and various external factors on the accuracy of the TDM-measured emissions. Factors tested include tracer location relative to the methane emissions hot spot of the landfill, distance from the landfill to the transect path, transect angle relative to the wind direction, and transect speed. Tracer location relative to the emissions hot spot is found to have the most significant effect on TDM accuracy, while transect angle relative to wind direction and transect vehicle speed are found to have negligible effects. The roles of wind direction and topography are also examined and found to have significant effects of the TDM’s accuracy.

In the second part of the dissertation, the same landfill area is simulated, and a surface flux parameterization is added to WRF to introduce wind-dependent variability to the land- fill emissions. Significant standard deviations were seen in the TDM-measured emissions in the previous chapter despite the prescribed landfill emissions remaining constant, and when TDM-measured emissions for steady and unsteady emissions simulations are compared, the variability in the TDM-measured emissions is found to be essentially the same even though the variability in true emissions is significantly different, pointing to possible errors inherent in the TDM’s ability to capture true emissions short-term variability. TDM-measured emissions standard deviation and TDM error are compared for eight different time periods over two different days to see whether different times of day and different atmospheric conditions affect the TDM. The smallest measured standard deviations and smallest errors are seen at night on both days, and measured standard deviation increases over the course of the day for both days, with the largest standard deviations seen in late afternoon shortly before sunset. TDM percent error does not exhibit a noticeable diurnal trend. Two different tracer configurations are used for the TDM simulations to obtain a range of standard deviations and percent errors for an optimal and less ideal tracer placement.

In the last part of the dissertation, a different landfill area is simulated and emphasis is placed on extensive comparison to field measurements. Four different days from different seasons are simulated to examine the seasonal and diurnal effects of wind-dependent variability on emissions. These simulations aim to help inform how limited measurement data can be used to extrapolate annual landfill methane emissions.