March-April 2009

Moving Up...to the Top of the Landfill

A field-validated, science-based model has been developed for the inventory of methane emissions in California landfills.

Tweet

California is typically at the forefront of innovative planning and regulatory strategies for environmental protection in the US. Two years ago, a research project was initiated by the California Energy Commission to develop an improved method for landfill methane emissions for the state greenhouse gas inventory. This article briefly describes the background and substance of this new methodology, which will be finalized during 2009 and is the topic of the keynote presentation at this year’s Solid Waste Association of North America (SWANA) Landfill Gas Symposium (the keynote presentation will be given on March 10, 2009, at the SWANA Landfill Gas Conference in Atlanta). This new methodology represents significant advances over previous strategies, as it incorporates site-specific information on the areal extent of cover materials (daily, intermediate, and final) and the effects of seasonal climatic variability on methane transport and emissions. Unlike previous methods, it does not rely primarily on a methane-generation model and thus moves up to the top of the landfill to directly address methane emissions. The last 10–15 years of research results on landfill methane transport, oxidation, and emissions have given us the knowledge and tools to develop a more science-based approach that also incorporates advances in modeling strategies.

No doubt about it: Methane is a potent greenhouse gas. In its Fourth Assessment Report (2007), The Intergovernmental Panel on Climate Change (IPCC) concluded that, on a 100-year time frame, each molecule of methane has a global warming potential 25 times higher than that associated with a molecule of carbon dioxide. For the US, the 2008 EPA report on national greenhouse gas emissions concluded that landfill methane was the second largest source of methane resulting from human activities. Landfill methane and other sources of biogenic methane (wetlands, ruminant animals, or rice production) result from the activity of microorganisms called “methanogens” which require strictly anaerobic conditions.

But there is also some good news: Globally, and for the US as a whole, landfill methane accounts for less than 2% of total annual greenhouse gas emissions (including carbon dioxide, methane, nitrous oxide, and fluorinated gases). Also, because of increased landfill gas recovery and use, landfill methane emissions have been declining in the US over the last decade. Because the atmospheric lifetime of methane is quite short (about a decade), reductions in methane emissions can also significantly reduce atmospheric concentrations over relatively short time frames. Thus, many countries are actively targeting reductions in methane emissions as an attractive mitigation strategy.

First, some historical background on landfill methane models and methods. More than 30 years ago, when the first commercial landfill gas recovery projects were being implemented, developers began using first-order kinetic models—in which the theoretical quantity of methane generated from the waste landfilled in a particular year is summed with the methane generation from waste landfilled in each previous year. However, there are often large uncertainties regarding waste quantities and composition, not to mention the assumptions for microbial methane generation rates from various waste fractions in a specific landfill. Thus the modeling must be adjusted after gas recovery has been implemented in order to more closely match reality.

These models were a good first step and are still extremely useful for commercial landfill gas recovery projects. However, for emissions specifically, we can now do much better using improved analytical tools, newer field-validated models, and reference data sets from a wide range of published field and laboratory studies. Ideally, modeling should be done in the context of a methane mass balance for a given site or cell, where the methane generated is partitioned into the methane recovered, emitted, oxidized, and migrated, as well as to changes in the quantity of methane stored within the landfill volume (see Fig. 1a). When there is good information available on waste quantities and composition, and field-validated models are applied, there can be a statistically significant relationship between the modeled generation and the actual recovery. For example, Fig. 1b plots landfill methane recovery versus landfill methane generation for seven full-scale landfill cells for which a field-validated multicomponent first-order model was used (without adjustment) as the basis for methane generation—note the good agreement between modeled methane generation and measured methane recovery.

However, Fig. 1c tells a different story for the relationship between modeled gas generation and the corresponding methane emissions. Here the same methane generation data from Fig. 1b are plotted against actual methane emissions, which were measured in the field using two different methods (tracer gas release and chamber methods). First of all, note that the methane emissions were relatively small but varied over several orders of magnitude. This range is consistent with worldwide literature indicating that rates can vary by six to seven orders of magnitude (in units of g m^-2 day^-1). Secondly, it is important to include methane oxidation in the “net” measured emissions. This is a second microbial process complementary to methane generation, which, in aerobic cover soils, can oxidize methane before it is emitted to the atmosphere, thus reducing emissions. According to the worldwide literature for landfill soils, as well as for forest, grassland, wetland, and other types of soils, rates for methane oxidation can also vary by several orders of magnitude in response to environmental variables. The specific consortium of methane-oxidizing microorganisms and their activity or rate at a specific location is a function of the field conditions (e.g., temperature, moisture conditions, or quantities of methane and oxygen available in the soil gas). In general, higher rates of methane oxidation occur over a relatively narrow temperature range (10–40°C; 50–104°F) and require a moderate level of soil moisture (but not too much). Therefore, in order to model methane oxidation at a specific location, the soil texture, the annual cycles for soil moisture and temperature, and the relationship between these variables and oxidation rates should be known. For the current project, these complex relationships were elucidated through reference to refereed literature as well as extensive customized laboratory incubation studies targeting California landfill soils.

In general, we need methodologies for landfill methane emissions that can realistically incorporate methane generation, transport, oxidation, and “net” emissions to the atmosphere. Internationally, one of the major activities of the IPCC is to develop national inventory methods for greenhouse gas emissions. Historically, most developed countries have used a first-order kinetic model for methane generation to estimate annual emissions from buried waste, and indeed, such models are the basis for the Tier 1 and Tier 2 default methods in the current IPCC guidelines. However, these guidelines also recognize the current state-of-the-science with respect to validated methods for field measurements, an expanded database of field measurements in the refereed literature, and evolving theoretical and empirical models appropriate for national or regional inventories. Thus, for the first time, these guidelines also permit higher tier methods based on field measurements and more advanced models. Given this flexibility and realizing the limitation of existing methods, the California Energy Commission’s Public Interest Energy Research (PIER) Program decided to fund this research project. This new method for California is consistent with Tier 4 methods under the 2006 UNFCCC guidelines and would be the first such regional inventory methodology based on site-specific landfill methane emissions.

So how does the new inventory methodology for California work? An overview of the methodology is given in Figure 2a. First, a minimal amount of site-specific information is annually entered into a user-friendly template, including the geographical location (latitude/longitude), the specific cover designs and areas (e.g. daily, intermediate, and final covers), the filling history, and the history of landfill gas recovery associated with various cover types (if applicable). Either the site-specific waste composition can be entered, or the user can access regional published data for California. Currently, the model can include up to 10 different cover designs at each site, either defaulting to a list of standard California designs or using a customized “cover designer” to set up cover profiles. For California, some of these data are linked to the existing Solid Waste Information System (SWIS) database to reduce data input requirements. Following data entry, a series of climatic and soils models are linked to yield total annual site emissions. Included are improved and updated versions of existing field-validated USDA models: a temperature model (GlobalTempSIM), solar radiation model (SolarCalc), precipitation model (GlobalRainSIM), and soil temperature/moisture model (STM2). (The sub-models are freely available from the US Department of Agriculture, Agricultural Research Service Web site, and the models have been described and validated in two-peer reviewed publications available upon request.) Then, once the annual cycle of soil temperature and moisture through each designated cover design is known for a specific location, a new field-validated methane transport and oxidation model is activated for each cover type at each site. This model utilizes a 1-D gas diffusion model (for oxygen and methane) with laboratory-derived kinetic determinations for methane oxidation as a function of both temperature and moisture to predict the rate of methane oxidation at various depths at each time step (currently five minutes). The final task is to sum the time-series data over a yearly cycle for each cover type to calculate the annual methane emissions and oxidation for an entire site.

The development of this methodology has included extensive field validation and supporting laboratory studies. Over a two-year period, seasonal field measurements were conducted at two sites in California using static chambers to measure the gas exchange at the surface. Despite known limitations of the chamber methodology, this is one of a few methodologies that is capable of providing a direct measurement of gaseous emissions from a well-defined area.

The static chambers consisted of two parts: a circular base assembly (12 inches high and 18 inches deep) that was inserted into the soil prior to measurement (approximately 5 inches) and a hemisphere-shaped top that exactly fit into a customized trough on top of the base. The trough was filled with distilled water (see Figure 2b), and hand clamps were also used to provide good sealing between the base and the top over short monitoring periods (30 minutes or less).

Periodic gas samples from the enclosed chamber were transferred to helium-flushed serum vials for shipment and subsequent analysis on a laboratory gas chromatography system for major gases, or on an isotope mass spectrometer for stable carbon isotopes of methane. Additionally, the GPS coordinates, air temperature, soil moisture, and chamber temperature were recorded at every chamber location. At many chamber locations, a 4-inch soil gas sample was also collected for determination of soil gas concentrations and the stable carbon isotopic composition of methane. In addition, semipermanent gas probes were also installed in the final and intermediate cover types at the two sites to collect information on the soil gas profile for methane and other gases through the cover materials. Weather stations were also deployed at each site (intermediate and final cover) for monitoring the soil temperature, soil moisture, air temperature, relative humidity, wind speed, solar radiation and precipitation at each site to validate the model predictions. These data were collected continuously during the two-year period where seasonal sampling provided surface methane emissions data from each cover type for the validation of the emission modeling. Isotopic ratios in the emitted methane, the soil gas methane, and the deep (anaerobic zone) methane were used to calculate the fractional methane oxidation through each cover profile. In addition, selected data from an ongoing Waste Management program to monitor methane emissions and oxidation at several California sites are also being used to provide additional field validation.

Figure 2b depicts the seasonal variability of emissions from three different cover types at one of the California sites. Each bar represents the average of all methane emission measurements from a respective cover plus the associated standard deviations. The daily cover consists of 12 inches of silty sand; the intermediate cover is 3 feet of silty sand soil; and the final cover is 7–9 feet of silty sand soil. All fluxes measured at this site were low (<1 g m-2 day-1). As can be seen in Figure 2b, there are significant fluctuations in the behavior of the intermediate cover over the different seasons (going from net methane emissions to net oxidation of atmospheric methane, e.g., negative fluxes). For this particular site, there were no statistically significant differences within the final cover or daily cover emission measurements among the seasonal samplings.

The dynamics of methane oxidation for the intermediate cover soil as a function of temperature and moisture are shown in Figure 2c. There is an optimum temperature between 15°C and 30°C and optimum moisture between 10% and 30 % (gravimetric moisture, dry basis). Moreover, predicted seasonality differences from the laboratory oxidation experiments were mirrored in the observed emissions. Cover soils during the initial field campaigns in central California in March 2007 and August 2007 were extremely dry. The total precipitation in 2007 was 11.3 centimeters, whereas in 2008 the total was 37.1 centimeters. Because of its dry state (<5% gravimetric moisture), the soil collected during the 2007 drought was not capable of oxidizing methane. In the laboratory, this soil had to be incubated with added moisture over a considerable period of time (days) to reestablish its oxidation capacity. This is due to the production of spores by the microbes (methanotrophs) once conditions become nonfavorable for growth.

Upon reestablishment of favorable conditions, the methanotrophic populations start to recover. Along with other, more direct responses to such environmental conditions as temperature and moisture, these lag periods are an important component in the modeling of microbial methane oxidation, especially for many parts of California where seasonally dry soils are the norm.

Some final words on project timing and implications relative to previous methods. Comparisons of modeled emissions to field data, model adjustments, and sensitivity analysis will be completed this year. The product of this project will be a Web-based, user-friendly methodology based on the extension, application, and linkage of previously field-validated modeling and measurement methods, regional soils and climatic databases, and new methane emissions modeling. Existing inventory methodologies rely on a first-order kinetic model for theoretical methane generation based on the mass of waste in place and the filling history. This is the first regional inventory methodology for landfill methane emissions that relies on field-validated modeling of emissions as “net” emissions (inclusive of methane oxidation) rather than estimating emissions strictly on methane-generation modeling.

This more direct approach has fewer uncertainties than previous approaches based on generation, can be directly field-validated, and, indeed, is being validated through a large number of field measurements, supporting laboratory studies, and reference to other research results during the last decade. The latest (2006) IPCC inventory guidelines permit the use of such higher tier methods based on historic field measurements and more advanced models; thus, this project for California will be the first regional inventory methodology for landfill methane utilizing these higher tier methods.

Author's Bio: G. Franco is with the California Energy Commission, Sacramento, CA.

Author's Bio: J.E. Bogner is principal with Landfills + Inc. in Wheaton, IL, and the University of Illinois Chicago.

Author's Bio: Daniel P. Duffy, P.E. is an environmental engineer for CEC Inc. in Cincinnati, OH.

Author's Bio: J. Chanton is with Florida State University at Tallahassee, FL.

Author's Bio: K. Spokas is with the US Department of Agriculture, Agricultural Research Service, in St. Paul, MN.