Grantee Research Project Results
2005 Progress Report: Investigation of the Interactions between Climate Change, Biomass, Forest Fires, and Air Quality with an Integrated Modeling Approach
EPA Grant Number: R832277Title: Investigation of the Interactions between Climate Change, Biomass, Forest Fires, and Air Quality with an Integrated Modeling Approach
Investigators: Shankar, Uma , Hanna, Adel , Fox, Douglas G. , Binkowski, Francis S. , Xiu, Aijun , Holland, Andy , Seppanen, Catherine , Mattocks, Craig , McNulty, Steve
Current Investigators: Shankar, Uma , Hanna, Adel , Fox, Douglas G. , Binkowski, Francis S. , Xiu, Aijun , Holland, Andy , Seppanen, Catherine , Vukovich, Jeff , McNulty, Steve
Institution: University of North Carolina at Chapel Hill , USDA
EPA Project Officer: Chung, Serena
Project Period: March 13, 2005 through March 12, 2008 (Extended to March 12, 2009)
Project Period Covered by this Report: March 13, 2005 through March 12, 2006
Project Amount: $726,566
RFA: Fire, Climate, and Air Quality (2004) RFA Text | Recipients Lists
Research Category: Climate Change , Air Quality and Air Toxics , Air
Objective:
Forest fires not only change landscapes and destroy property but also emit trace gases and aerosols (e.g., CO, CH4, NOx, and black carbon) that affect regional and global air quality, with consequences to human health and climate through their interactions with solar radiation. These impacts can be felt over long distances because of the long-range transport of these pollutants both as primarily emitted species and as precursors for other pollutants formed in the atmosphere through photochemical reactions. Recently, the increased frequency of large fires in the United States has been thought to be associated with short-term changes in climate variables such as precipitation and temperature that have exacerbated the conditions for fire occurrence. Thus, the overall goals of this research project are to: (1) assess the impact of climate change and variability on biomass and forest fires; (2) evaluate the impact of the evolving emissions from forest fires on ozone and particulate matter air quality; and (3) determine the regional climate response to air quality changes in the Southern United States using an integrated modeling system. Integration of some of these feedbacks in the simulations will allow a more faithful representation of future-year emissions from both biogenic and fire sources, and their impact on air quality, and can thus help assess the air quality and ecosystem benefits of various fire management scenarios, currently not included in most emissions control evaluations.
The specific objectives of this research project are to: (1) investigate the impacts of climate change on vegetative cover and fuel characteristics, the consequences for fire frequency and intensity, and feedbacks to biomass load and biogenic emissions under managed and wildfire scenarios; (2) examine changes in regional air quality caused by the evolution of anthropogenic and biogenic emissions in response to various fire scenarios; and (3) investigate the feedback of the air quality changes to regional climate variables.
Progress Summary:
The project was a collaboration between researchers at the University of North Carolina at Chapel Hill (UNC); Dr. Douglas Fox, a private consultant and fire modeling expert; and researchers at the U.S. Department of Agriculture Forest Service (USFS) Southern Global Change Program (SGCP) led by Dr. Steven McNulty. The modeling system for the studies listed in the project objectives is shown in Figure 1. Research in Year 1 of the project focused on three key areas of development and implementation within this system: the forest growth model, Photosynthetic (nitrogen) Evapotranspiration (PnET) (Aber, et al., 1996); the fire emissions model, BlueSky-EM (Sestak, et al., 2002); and the coupled meteorology-chemistry model, METCHEM. At the start of the project, UNC submitted and got U.S. Environmental Protection Agency (EPA) approval of its Quality Management Project Plan (QMPP), delineating the quality assurance roles of project personnel. In July 2005, UNC organized a 2-day project meeting, including seminars by Dr. Fox and Dr. McNulty, to review the QMPP, understand the modeling issues, and scope out tasks described in the next sections.
Figure 1. Schematic of the Modeling System. Year 1 development areas are highlighted.
Forest Process Modeling
The impacts of climate change on forest biomass production are being simulated using the PnET model. During Year 1 of the project in collaboration with Dr. McNulty’s group, we studied the VisualBasic version of the PnET model used in the SGCP applications for its portability to the UNC computational platform. After a detailed investigation of the model structure and input databases, we determined that the number and length of simulations required by the project called for a more efficient and flexible model version than this version, which also features the use of large MySQL input databases that are not ideal for rapid data transfer. After investigating the various research and application versions used around the country, a C++ version from the University of Minnesota was chosen as the working version because of its computational speed and flexibility. It has been ported to the UNC computational platform and used as the basis to develop a Java version for increased maintainability and ease of use. The model code has been enhanced to use input/output (I/O) netCDF files, as these are self-describing, are machine-independent, provide fast direct-access data retrieval, are independent of a data server, and are in the I/O format of the other major modeling system components. PnET’s I/O also has been enhanced to read four key meteorological inputs, namely monthly average precipitation, minimum and maximum temperature, and solar radiation from the Community Climate System Model (CCSM), which can be tailored to reflect various climate change scenarios defined by the Intergovernment Panel on Climate Change (e.g., caused by CO2 doubling or a CO2 mitigation policy). This promotes consistency within the system, as CCSM also provides initial and boundary conditions to METCHEM and is used to simulate air quality and its climate feedbacks. We developed a method based upon Iqbal (1983) to calculate Photosynthetically Active Radiation (PAR) input to PnET in number of incident photons per unit area-per unit time from the CCSM output of downwelling solar flux (Watt/m2). Rainfall rate input to PnET is derived from CCSM output as the total of convective and large-scale precipitation rates. We ran test simulations of the Java code and were able to replicate 10-year Harvard Forest benchmark simulation results for biomass components and soil respiration to within machine precision, at a speed comparable to the C++ version of the model (Figure 2).
(a) | (b) |
Figure 2. Comparisons of Forest Carbon Production Per Unit Area-Per-Year for Various PnET Components Between the C++ Version (C) and the Java Version (J): (a) Daytime, Nighttime, and Total Respiration; (b) Foliar, Wood, Root, and Total Primary Production
Fire Emissions Modeling
In this project period, we researched the ideal model for the generation of fire emissions. Initial plans were to use the Community Smoke Emissions Model (CSEM) as it was designed specifically to provide historical fire emissions estimates for regional-scale air quality modeling using fuel loads from the National Fire Danger Rating System (NFDRS), along with historic fire data from an inventory of fire activity. An alternative model under consideration was the BlueSky smoke emissions model more recently developed by USFS for real-time applications, based upon many of the CSEM algorithms. One advantage of CSEM over BlueSky was that it could read input meteorological data from the Fifth Generation National Center for Atmospheric Research/Penn State Mesoscale Meteorological Model (MM5) and could calculate plume rise using the Briggs algorithm to vertically allocate the fire emissions (this capability was not available in BlueSky). Through a supporting project with the National Oceanic and Atmospheric Administration (NOAA), however, BlueSky has been linked to the Sparse Matrix Operations Kernel Emissions (SMOKE) model. This linked BlueSky-EM model can be used without the need to read hourly meteorological data; SMOKE performs this function, and calculates plume rise for smoke from fires and power plant plumes. BlueSky-EM has been enhanced to use fuel load data from one of three databases: the Fuel Characteristics Classification System (FCCS) developed by the Pacific Northwest Research Station of the USFS , the NFDRS, and the Hardy database developed for Western U.S. applications at fine resolution (Hardy, et al., 2000). Recent improvements to FCCS have expanded spatial coverage to all of the contiguous United States at 1-km spatial resolution, with 285 fuel bed classes; this provides considerably more detail than the NFDRS. We compared the three databases and selected FCCS as the default for our applications, because of its detailed spatial heterogeneity, large coverage area, and compatibility with the resolution of the Biogenic Land Cover Database (BELD3) used by the Biogenic Emissions Inventory System Version 3.
We investigated and identified land use types appropriate to represent the burned land in the BELD3 landcover data in consultation with Dr. Tom Pierce of NOAA who was the Project Officer of a supporting project to investigate the impacts of wildland fire emissions on real-time air quality forecasts. We researched the simulation of future fires and the development of a stochastic model for providing fire activity data to BlueSky-EM for future years and made contact developers of the Fire Scenario Builder (FSB) (McKenzie, et al., 2006) at the Pacific Northwest Research Station of USFS. FSB uses lightning as the trigger to stochastically model future fires; it has been linked directly to FCCS and BlueSky and tested over the Northwest in the 2003 season. We studied the latest literature on this model and also made contact with researchers at the Southern Research Station of USFS in Research Triangle Park, North Carolina, to discuss the considerations for fire triggers specific to the Southern United States, where arson and other human factors are much more frequent ignition mechanisms (Mercer and Prestemon, 2005). Accordingly, we arranged for these two groups to give seminars and attend a meeting at UNC in September 2006 to provide guidance on adapting FSB for our Southern U.S. applications.
Coupled Meteorology-Chemistry Model Enhancements
Under a previous Science To Achieve Results (STAR) grant, UNC investigators developed an air quality modeling system with integrated meteorology and chemistry (METCHEM) as a modular, physically and numerically consistent, fully integrated regional-scale atmospheric dynamics and chemistry modeling system. It is based on further development and refinement of three existing models: the MM5, the Multiscale Air Quality Simulation Platform (MAQSIP) (Mathur, et al., 2005), and the SMOKE modeling system. The integrated modeling framework enables investigation of the feedbacks of radiatively important trace species to the atmospheric dynamics. METCHEM applications over the Eastern United States have been able to capture the increase in optical depth because of the presence of aerosols, resulting in a reduction in the shortwave radiation reaching the ground and consequently, in the surface temperature and planetary boundary layer height.
In Year 1 of the project, we made enhancements to METCHEM’s radiative transfer model to speed up considerably the calculation of aerosol optical properties. We have updated the refractive index estimates for various classes of aerosols (water soluble, insoluble, elemental carbon, and water) using the Optical Properties of Aerosols and Clouds software package of Hess, et al. (1998). Scattering and absorption cross-sections for aerosols and gases and predicted aerosol size distribution parameters (number concentrations and mode mean diameters) from MAQSIP are used in our fast optics code based upon analytical expressions following Heintzenberg and Baker (1976) and Willecke and Brock (1977) to calculate the single scattering albedo, asymmetry factor, and aerosol optical depths (AODs). Our initial test simulations with METCHEM following these enhancements found AOD estimates for regions in the model domain with high relative humidity (RH) to be unrealistically high. After identifying and correcting various coding errors in the radiative transfer model, we traced the main source of the persistently high AODs back to very high aerosol liquid water content (LWC, greater than 1,000 µg/m3) predicted by the ISORROPIA thermodynamic model in regions with high inorganic fractions and RH above 95 percent (e.g., in the Eastern United States in summer). This led to our setting an upper limit of 95 percent for the RH input to ISORROPIA and resulted in much lower LWCs and AODs. This was an important finding that we communicated to EPA scientists, who have since corrected the same problem identified in the Models-3 Community Multiscale Air Quality (CMAQ) model. Their tests of CMAQ with the RH correction for January and July 2001 showed significant impacts on the aerosol composition (increased SO4 and reduced NO3), reduction in the particle size, and better agreement of model results with observations (Bhave, private communication).
We defined the METCHEM simulation domains based upon extensive discussions with SGCP and Dr. Fox. The model will be run over a North American domain (Figure 3) using fire emissions inputs from BlueSky-EM processed by SMOKE along with other inventoried emissions for 2002 (the base year). As forest inventory inputs for the Southern United States currently obtained from the Forest Inventory Analysis system database are not available yet with comparable quality for the rest of the country, this simulation would not include PnET output. It would provide boundary conditions, including the effects of large Western U.S. fires, to a nested Southern U.S. domain shown in Figure 3, an are where PnET has been evaluate extensively and that would include PnET results. We also defined two 6-week simulation periods in summer and fall, respectively representing the seasons for catastrophic wild fires and prescribed burns in 2002, 2015, 2030, and 2050. Future year simulations will consider the Intergovernmental Panel on Climate Change scenarios (Business-as-Usual and CO2 mitigation policy) in the CCSM initial and boundary inputs to METCHEM.
Figure 3. METCHEM Simulation Domains; Domain D02 will Include the Impacts of PnET-Predicted Biomass Changes on the Fire Emissions Inputs
Future Activities:
In Year 2 of the project, the main objective will be completion of the modeling system implementation, and applications of the system to the base year and a future year to examine the fire emissions impacts on air quality under various fire scenarios (wild fires only versus National Fire Management Plan). The project tasks will focus on the following areas:
- Link PnET model output to the BELD3 to reflect changes in biogenic emissions resulting from biomass changes.
- Reflect changes caused by fires in the BELD3 land cover data.
- Adapt the FSB developed by USFS to include fire ignition scenarios prevalent in the Southeastern United States (e.g., arson).
- Perform simulations for the continental United States at 108 km, nesting down to the Southeastern United States at 36 km for two seasonally representative 6-week periods in 2002.
- Evaluate the 2002 simulations against available observations.
- Complete 2015 METCHEM simulations for two seasonally representative periods for wild fires only and implementation of the National Fire Plan.
- Analyze 2015 simulations for aerosol optical properties and radiative impacts and compare with 2002 results.
- Prepare annual project report.
References:
Aber JD, Ollinger SV, Driscoll CT. Modeling nitrogen deposition in forest ecosystems in response to land use and atmospheric deposition. Ecology Modelling 1997;101:61-78.
Hardy C, Burgan RE, Ottmar RD, Deeming JE. A database for spatial assessments of fire characteristics, fuel profiles, and PM10 emissions. Journal of Sustainable Forestry 2000;11:DOI:10.1300/J091v11n01_09, 229-244.
Heintzenberg J, Baker M. Spherical particle populations: approximate analytic relationship between size distribution parameters and integral optical properties. Applied Optics 1976;15(5):1178-1181.
Hess M, Koepke P, Schult I. Optical properties of aerosols and clouds, 1998: the software package OPAC. Bulletin of the American Meteorological Society 1998;79(5):831-843.
Iqbal M. An Introduction to Solar Radiation. New York, NY: Academic Press, 1983:390.
Mathur R, Shankar U, Hanna AF, Odman MT, McHenry JN, Coats CJ, Alapaty K, Xiu AJ, Arunachalam S, Olerud DT, Byun DW, Schere KL, Binkowski FS, Ching JKS, Dennis RL, Pierce TE, Pleim JE, Roselle SJ, Young JO. Multiscale air quality simulation platform (MAQSIP): Initial applications and performance for tropospheric ozone and particulate matter. Journal of Geophysical Research-Atmospheres 2005;110(D13):1535-1553.
McKenzie DM, O’Neill SM, Larkin NK, Norheim RA. Integrating models to predict regional haze from wildland fire. Ecology Modelling (in press, 2006).
Mercer DE, Prestemon JP. Comparing production function models for wildfire risk analysis in the wildland–urban interface. Forest Policy and Economics 2005;7:782-795.
Sandberg DV, Peterson JL. A source strength model for prescribed fire in coniferous logging slash. Presented at the Annual Meeting of the Air Pollution Control Association, Pacific Northwest Section, Portland, OR, 1984.
Sestak M, O’Neill S, Ferguson S, Ching J, Fox D. Integration of wildfire emissions into models-3/CMAQ with the prototypes community smoke emissions modeling system (CSEM) and BlueSky. In: Proceedings of the Second Annual Community Modeling and Analysis Workshop, Research Triangle Park, NC, October 21-23, 2002.
Willeke K, Brockmann JE. Extinction coefficients for multimodal atmospheric particle size distributions. Atmospheric Environment 1977;11:995-999.
Wythers KR, Reich PB, Tjoelker MG, Bolstad PB. Foliar respiration acclimation to temperature and temperature variable Q10 alter ecosystem carbon balance. Global Change Biology 2005;11:435-449.
Journal Articles on this Report : 1 Displayed | Download in RIS Format
Other project views: | All 21 publications | 1 publications in selected types | All 1 journal articles |
---|
Type | Citation | ||
---|---|---|---|
|
McKenzie D, O’Neill SM, Larkin NK, Norheim RA. Integrating models to predict regional haze from wildland fire. Ecological Modelling 2006;199(3):278-288. |
R832277 (2005) R830962 (2006) R830962 (Final) |
Exit Exit Exit |
Supplemental Keywords:
forest biomass, fire emissions, land cover changes, air quality climate feedbacks, meteorology, landuse, aerosols, global climate change, climate models, atmospheric chemistry, environmental stress,, RFA, Scientific Discipline, Air, Ecosystem Protection/Environmental Exposure & Risk, Chemistry, climate change, Air Pollution Effects, Monitoring/Modeling, Environmental Monitoring, Ecological Risk Assessment, Atmosphere, Community Smoke Emissions Model, anthropogenic stress, environmental measurement, meteorology, climatic influence, global ciruclation model, global change, ozone depletion, air quality model, biomass, climate models, terrestial ecosystem model, environmental stress, coastal ecosystems, ecological models, climate model, forest resources, Global Climate Change, atmospheric chemistry, climate variabilityRelevant Websites:
http://cf.unc.edu/cep/empd/projects/FIRE/ Exit
http://www.cep.unc.edu/empd/projects/integrated/ Exit
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.