The tools used to estimate landfill gas production need better science.
Despite continued trends toward diversion, about 69 percent of the 389 million tons of municipal solid waste (MSW) generated in the United States are landfilled according to the most recent State of Garbage Report published by Biocycle and Columbia University. Compare that to 30 years ago when nearly 90 percent of MSW was landfilled and it becomes clear that, historically, the vast majority of U.S. MSW still today resides in landfills. Microorganisms decompose a substantial portion of the organic matter in landfilled waste into gas that is roughly equal volumes of methane and carbon dioxide.
This landfill gas – more specifically the methane portion – is now being utilized beneficially to create heat, electricity and, more recently, compressed or liquefied natural gas. Thus, even if materials stopped going to landfills tomorrow, there is still a lot of potential in the use of landfill gas as an alternative energy source.
A model is typically employed to evaluate the production potential of landfill gas. Traditionally, EPA’s Landfill Gas Emissions Model, or LandGEM, has been the most widely used for this purpose. The LandGEM model is relatively simplistic and contains two key parameters that essentially determine how much landfill gas is produced over a specified time frame.
One of these parameters is the methane production potential of the waste, or “Lo,” and is based on waste composition, or more specifically, the relative fraction of organics in the waste. Since microbes are responsible for landfill gas generation, Lo is analogous to how much food is available to feed them. More degradable organics in the waste equates to more available food, which results in a higher Lo, while less organics yield a lower Lo.
The second parameter is the decay rate, or "k", which is a function of the speed at which the waste decomposes. The decay rate is analogous to how quickly the microbes eat their food and is dependent on many factors, including moisture content, temperature, microbial populations, pH, and spatial factors that enhance or reduce the contact between organics material, moisture and microbes.
The values used in the LandGEM model for these two parameters, Lo and k, were determined nearly 20 years ago and are based upon landfill management practices used in the early 1990s. Much has changed since that time. We have more and better data on waste composition and how much methane can be generated from specific waste components, both of which affect Lo. The understanding of landfill gas collection and collection efficiency has improved. Additionally, landfill operations also are different today and relatively new practices such as leachate recirculation have a significant impact on how quickly waste degrades, which in turn affects the k value.
EPA knows the current parameters need to be updated. A statement from their AP-42 database reads “Although the recommended default k and Lo are based upon the best fit to 21 different landfills, the predicted methane emissions ranged from 38 to 492 percent of actual…”. It’s this variability from LandGEM that have led some to share with me that while LandGEM is used to provide a first estimate of methane production, many subsequently employ a substantial amount of empirical experience, best professional judgment and “fudge” factoring to predict how much landfill gas might be expected from a landfill.
These issues demonstrate the need for better science in landfill gas estimation. Given that landfill owners and landfill gas field developers have placed substantial investments in collection and landfill gas to energy systems, more accurate determinations during initial evaluations would reduce financial risk, improve the utilization of landfill gas as an alternative energy source and have the potential to increase profitability. The importance of this has become more acute over the years as smaller landfills or landfills with lower gas production have been considered as candidates for the implementation of gas-to-energy systems. In these situations, there is a much smaller margin of error.
In part 2 of this series, I’ll go over some of the new developments that are addressing these issues on the research and data development front.