Grantee Research Project Results
Final Report: Atmospheric Ammonia Emissions from the Livestock Sector: Development and Evaluation of a Process-based Modeling Approach
EPA Grant Number: R834549Title: Atmospheric Ammonia Emissions from the Livestock Sector: Development and Evaluation of a Process-based Modeling Approach
Investigators: Adams, Peter
Institution: Carnegie Mellon University
EPA Project Officer: Chung, Serena
Project Period: May 1, 2010 through April 30, 2014 (Extended to November 30, 2015)
Project Amount: $483,827
RFA: Novel Approaches to Improving Air Pollution Emissions Information (2009) RFA Text | Recipients Lists
Research Category: Air Quality and Air Toxics , Air
Objective:
We propose multi-faceted research to enhance our understanding of NH3 emissions from livestock feeding operations. A process-based emissions modeling approach will be used, and we will investigate ammonia emissions from the scale of the individual farm out to impacts on regional nitrogen deposition and particulate matter formation. Proposed tasks are: (a) develop Farm Emission Models (FEMs) and National Practice Models (NPMs) for beef cows, swine, and chickens; (b) evaluate the FEMs against unprecedented emissions measurements made during the NAEMS study; (c) use the FEMs and NPMs to generate a process-based NH3 emissions inventory for the United States; (d) incorporate revised livestock emissions into the CMU Ammonia Emissions Inventory; (e) evaluate the performance of the revised inventory in a chemical transport model (CTM) against a suite of ambient observations; (f) perform policy-relevant chemical transport modeling to understand the sensitivity of inorganic PM2.5 to SO2, NOx, and NH3 under current and future regulatory regimes; and (g) disseminate the revised inventory to the public via the World Wide Web.
The objective of this work is to develop a set of emissions modeling tools that predict the seasonality and amount of ammonia emissions from livestock operations. These emissions contribute to ambient PM2.5 concentrations when ammonia reacts with nitric acid gas to form ammonium nitrate particulate matter. Ammonia emissions are highly variable in space and time depending on local climate conditions and farming practices. Ammonia emissions models should adequately account for the major regional and seasonal variations in livestock emission factors.
Summary/Accomplishments (Outputs/Outcomes):
Project Activities
The original proposal outlined a research plan consisting of five tasks. Here, we summarize achievements for each of these tasks.
Task 1: Development of Farm Emissions Models (FEMs) for Major Livestock Types
This task proceeded as outlined in the proposal and the results were published in Atmospheric Environment in 2015 (McQuilling and Adams, 2015).
Briefly, we constructed FEMs for beef cattle, swine, and poultry that are based on mass balance principles and included separate submodels for each of the stages of the manure management system: housing, storage, application, and grazing. The FEMs were based on a review of academic literature and farming manuals to include the relevant management practices for each livestock type. In the review, we compiled a database of measured emission factors at different farms and a number of explanatory parameters: temperature, wind speed, pH, feed or manure nitrogen content, etc. Separate FEMs for each livestock type (beef, poultry, etc.), manure management stage (housing, storage, etc.) and major practice category (e.g., deep pit vs shallow pit swine housing) were tuned to measurements using mass transfer resistances as tunable parameters to capture how emissions depend on temperature, wind speed, season, etc. In some cases, there was insufficient data for some ammonia sources (e.g. beef cows on pasture), so that best guess values had to be used rather than tuning full FEMs.
Task 2: Evaluation of Farm Modules vs. NAEMS
This task also proceeded as outlined in the proposal. A manuscript detailing this work is in the final stages of revision and will be submitted for publication in the next ~ 2 months.
The National Air Emission Monitoring Study (NAEMS) provided a unique opportunity to enhance and evaluate the FEMs developed in Task 1. Unique features of these measurements were: (a) the number of farms included; (b) the length of the measurements (> 1 year); (c) consistent methodologies applied; and (d) the measurement of a number of explanatory variables (e.g., manure nitrogen) not always available in the literature.
After reviewing NAEMS results, it was apparent that there were some systematic differences with prior literature. Notably, as measurements were made year-round, NAEMS included emissions measurements at a wider temperature range (especially cold temperatures) than were available previously. This necessitated re-calibration of the FEMs developed in Task 1 since extrapolation of the prior fits to new conditions yielded poor results.
Using the NAEMS data, we evaluated the model’s capability to predict seasonal and daily variability. Except for broiler housing, the seasonal r2 values for FEMs vs NAEMS farms are in the 0.32-0.63 range, indicating that the FEMs capture much of the seasonal variability. Although the FEMs were originally developed to capture seasonal cycles, we also examined their ability to predict daily variations in ammonia emissions. R2 values in this case were lower than for seasonal variability, but 9 out of 15 farms showed modest predictive power for daily variability (r > 0.25).
Task 3 (Development of National Practice Models for Major Livestock Types) and Task 4 (Updates to the CMU Ammonia Inventory)
These two tasks are closely related, so we summarized them together. Different than our prior work (Pinder et al., 2004), we were unable to obtain the raw USDA survey data from the National Animal Health Monitoring System (NAHMS) due to increased concerns over data confidentiality. Therefore, instead of building our own statistical models based on the raw data, we used regional summaries of farming and manure management practices provided to us by USDA analysts.
In her thesis work, Dr. McQuilling created a national inventory for livestock ammonia emissions for 2011. Subsequently, she built another inventory for 2014 using the same methods, which has been provided to EPA’s National Emissions Inventory (NEI) team. Briefly, FEMs for beef, dairy, swine, and poultry are run multiple times, and the results combined to create the inventory. Meteorological inputs of temperature, wind speed, and precipitation are taken from daily values provided by the National Climate Data Center (NCDC). Farming practices are from USDA NAHMS as mentioned above. FEMs are run for each combination of livestock type, NCDC climate division, and manure management practice to predict daily emission factors from a full suite of possible farms. For each livestock type, these are then weighted by the prevalence of those farming practices in any given county to estimate county-average emission factors and multiplied by animal population numbers to provide total emissions.
Compared to EPA’s NEI2011v2, our total annual emissions differ by only ~ 10%. However, the share of emissions from swine production is higher in our results compared to NEI while dairy is much lower. These swine differences stem ultimately from the NAEMS measured emission factors, which for example, were significantly higher from manure storage at swine operations during summer than much of the previous literature. Additionally, our seasonal cycle is stronger than the one we had used in previous work (Pinder et al., 2006) with lower wintertime ammonia emissions, which we would expect to result in less wintertime PM nitrate and lower annual-average PM2.5 concentrations overall (Ansari and Pandis, 1998).
Further details about the methodology and data sources are provided in the Ph.D. thesis of Alyssa McQuilling. The relevant portion of this document is Chapter 4, entitled “Developing a new national inventory: data sources, management practices, and NH3 emissions.”
Task 5: CTM Modeling: Inventory Evaluation and PM2.5 Sensitivity
CTM modeling with this inventory was not performed during this work.
Technical Aspects
Our experience building a national ammonia emissions inventory for livestock using process-based farm models leads to several conclusions. Not surprisingly, estimation of emissions from outdoor sources (e.g., storage lagoons) is more challenging than enclosed sources because it requires estimating the amount of atmospheric dispersion and dilution between the source and the concentration measurement. Different methodologies (e.g., radial plume mapping vs. backwards Langrangian stochastic) can lead to different emissions estimates, especially when looking at daily variations. In other contexts, tracer release studies have helped to eliminate this ambiguity and may be promising here. They are, however, cumbersome.
Tuning emissions models to literature measurements or evaluating models versus those measurements is challenging when studies are sometimes inconsistent and incomplete in what they report about conditions. Meteorological conditions, feed nitrogen, manure nitrogen, and manure pH are all variables that have significant effects on emissions but are not always reported with measured emissions factors. Measured emission factors are considerably more valuable when these explanatory variables are also reported, and this should become the norm for future measurement studies.
The seasonal cycle of emissions, especially wintertime levels, is very important to predicting PM2.5 impacts even though wintertime emissions tend to be lower than summertime values. Prior to NAEMS, there was relatively little data available for emissions factors at colder temperatures. Similarly, there are some highly distributed emissions at low rates (e.g., beef cows on pasture) that may add up to be significant to the national inventory. There is frequently a lack of data on these, perhaps because much of the work has focused on local nuisance and odor issues.
Lastly, construction of a useful national inventory with regional and seasonal variability depends on knowledge of manure management practices – sometimes responsible for more uncertainty than the measured emissions factors themselves (Pinder et al., 2006). These data are frequently lacking or are difficult to obtain because of concerns about confidentiality.
The NAEMS study was a significant improvement on many of these fronts, especially because it provided long-term measurements of the seasonal cycle at many farms, filling in data gaps at lower wintertime temperatures, and also because it was more thorough and consistent about providing information (e.g., manure nitrogen) to explain observed variability in emissions.
Ultimately, successful construction and application of an inventory depends on input from a number of research communities: those making emissions factor measurements, experts on animal health and manure management practices, emissions modelers, and air quality modelers. Each of these groups has different perspectives, needs, and priorities. Collaboration between these communities (e.g., between EPA and USDA’s Agricultural Air Quality Task Force [AAQTF]), should be fostered and continued.
Conclusions:
Technical Effectiveness and Economic Feasibility
This study has demonstrated the feasibility of building a national emissions inventory for livestock ammonia using a process-based emissions modeling approach. To our knowledge, this is the first time that this has been done nationally for all livestock types. A national inventory has been developed with this approach, the results have been provided to the EPA’s NEI team, and incorporated into the NEI 2014 v1 emissions inventory. This makes EPA responsive to a finding by the National Research Council that “EPA and USDA should use process-based mathematical models with mass balance constraints for nitrogen-containing compounds, methane, and hydrogen sulfide to identify, estimate, and guide management changes that decrease emissions for regulatory and management programs” (National Academy of Sciences, 2003).
Environmental Benefits
The process-based modeling approach has several advantages compared to the traditional approaches of applying static emissions factors and/or emissions factor functions that are statistical fits. First, this approach naturally accounts for regional and seasonal variability that results from meteorological and manure management differences. Second, it leverages existing emissions factor measurements, puts them in context, and then uses a physically realistic framework to interpolate to other conditions. Third, because it incorporates a mass balance, it will not predict nitrogen emissions that exceed manure nitrogen levels. Fourth, it provides a physical framework to account for interactions between different stages of manure management operations (e.g., lower emissions from animal housing may result in higher manure nitrogen and emissions from manure storage and application or vice versa). Fifth, it provides a self-consistent and physically realistic framework to ask “what if” questions about the benefits of control strategies (e.g., changing feed nitrogen or manure management practices at any stage of the manure management train).
References:
Ansari, A.S. and S.N. Pandis, Response of inorganic PM to precursor concentrations. Environmental Science & Technology, 1998. 32(18): p. 2706-2714.
McQuilling AM, Adams PJ. Evaluating a process-based model for livestock ammonia emissions using long-term emisions measurements. (in preparation)
McQuilling, A.M. and P.J. Adams, Semi-empirical process-based models for ammonia emissions from beef, swine, and poultry operations in the United States. Atmospheric Environment, 2015. 120: p. 127-136.
National Academy of Sciences, Air emissions from animal feeding operations: current knowledge, future needs. 2003, National Academy of Sciences: Washington DC.
Pinder, R.W., et al., A temporally and spatially resolved ammonia emission inventory for dairy cows in the United States. Atmospheric Environment, 2004. 38(23): p. 3747-3756.
Pinder, R.W., et al., Temporally resolved ammonia emission inventories: Current estimates, evaluation tools, and measurement needs. Journal of Geophysical Research-Atmospheres, 2006. 111(D16).
Journal Articles on this Report : 1 Displayed | Download in RIS Format
Other project views: | All 3 publications | 1 publications in selected types | All 1 journal articles |
---|
Type | Citation | ||
---|---|---|---|
|
McQuilling AM, Adams PJ. Semi-empirical process-based models for ammonia emissions from beef, swine, and poultry operations in the United States. Atmospheric Environment 2015;120:127-136. |
R834549 (2013) R834549 (2014) R834549 (Final) |
Exit Exit Exit |
Supplemental Keywords:
Acid deposition, particulate matter, ecosystem, agriculture, ammonia emissions, livestock feeding operations;Progress and Final Reports:
Original AbstractThe perspectives, information and conclusions conveyed in research project abstracts, progress reports, final reports, journal abstracts and journal publications convey the viewpoints of the principal investigator and may not represent the views and policies of ORD and EPA. Conclusions drawn by the principal investigators have not been reviewed by the Agency.
Project Research Results
- 2014 Progress Report
- 2013 Progress Report
- 2012 Progress Report
- 2011 Progress Report
- 2010 Progress Report
- Original Abstract
1 journal articles for this project