Grantee Research Project Results
Final 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 , 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 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, as well as to climate through their interactions with solar radiation. Recent increases in wildfire frequency in the United States are thought to be associated with short-term changes in climate variables (precipitation, temperature) that have exacerbated fire weather. Thus, the goal of this research is to assess the impact of climate change on forest biomass and fires, the impact of the evolving fire emissions on air quality, and its regional climate impacts in the Southeastern United States using an integrated modeling system. Simulating some of these feedbacks would allow a more realistic representation of future-year biogenic and fire emissions impacts on U.S. air quality, and facilitate assessment of the ecosystem benefits of various fire management scenarios not currently considered in most emissions control evaluations. The following three primary objectives support the overall goal:
1. To 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. To examine changes in regional air quality due to the evolution of anthropogenic and biogenic emissions in response to various fire scenarios.
3. To investigate the feedback of the air quality changes to regional climate variables.
Summary/Accomplishments (Outputs/Outcomes):
The project was a collaboration between UNC, Dr. Douglas Fox, a fire modeling expert, and the USFS Southern Global Change Program (SGCP), led by Dr. Steve McNulty. In the second project year, Drs. Don McKenzie and Jeff Prestemon from USFS also became active consultants in modeling of future-year fires, and in accounting for human-induced ignitions as often occur in the Southeastern United States (Mercer and Prestemon, 2005) in addition to lightning-induced fires.
The modeling system for the studies listed in the project objectives is shown in Figure 1. The modeling components and associated modeling activities are described in the following sections.
Figure 1. Schematic of the modeling system.
Forest Process Modeling
Research in the first and second years focused mainly on compiling fuel data inferred from the applications of the Photosynthetic Evapo-transpiration Model (PnET), a detailed forest process model (Aber et al., 1996), over the selected modeling domain. The project thoroughly evaluated a few versions of PnET for its appropriateness in these applications. In Year 1 of the project, a daily plot-level PnET model version was investigated in consultation with the developers at the University of New Hampshire, and optimized to get the speedup needed for application to a large region. During Year 2, however, the actual application of this version was found to be unfeasible due to the large number of input parameters required for all the plots needed to model the Southeastern United States, and thus the monthly-average model version used by SGCP was chosen. Several vegetation and site-specific inputs are required to simulate plot-level forest growth besides climate parameters; plot-level data are then expanded to the county level using expansion factors compiled by the USFS for the various vegetation types. In Year 2, input data were collected for the base model year (2002) for the 13 states shown in Figure 2 from the Forest Inventory Analysis (FIA) databases in Knoxville, TN, and were quality-assured by comparing the model output against SGCP’s archived outputs for reasonable performance. Multivariate regression analyses using temperature, precipitation, and total mean solar radiation as the regression variables examined the correlation between biomass predicted for the base year and the forest fuel load data from the FIA, to infer future-year fuel loads from predicted live biomass. The fuel loads in the 13 states shown in Figure 2 were found to be very poorly correlated with the live biomass from PnET in the 2002 regression analyses, and thus the USFS recommended changing only the canopy fuel in the projections, with an assumed percent of canopy biomass apportioned to the canopy fuel. Estimates of 10-hr and 100-hr fuel were made using this approach, and other fuel classes were assumed constant in future years. In Years 2 and 3 of the project, data linkages were developed to provide future-year fuel loads to the USFS BlueSky-EM smoke emissions model (Sestak et al., 2002) to estimate future-year fire emissions.
Figure 2. Modeling domain for the PnET forest growth model application
Anthropogenic and Fire Emissions Modeling in 2002 and Future Years
In Years 1 and 2 of the project, 2002 NEI data were processed into model-ready emissions inputs for the model domains shown in Figure 3; at 36-km resolution over the continental United States for all emissions sectors, and at 12-km resolution for all but point-source (non-agricultural) fire emissions over the Southeastern United States. BlueSky-EM was tested, and scripts were developed to run it over the entire Southeast. Due to the complexity of the air quality modeling system, continuous modeling from present to 2050 is not feasible, and thus roughly decadal snap shots of the future are simulated. The method for projecting anthropogenic emissions for point sources to these selected years was investigated in Year 3 in consultation with the EPA Office of Air Quality Planning and Standards. Based upon the projection data available from EPA, a change was made in the future years selected for the modeling to include 2020 rather than 2015, along with the years 2030 and 2050. During the no-cost extension in Year 4, work began to process emissions from non-fire sectors over the Southeastern United States and to merge them with fire emissions.
Also in Years 3 and 4, both lightning and human ignition probabilities were incorporated in a statistical model of fire activity (areas burned) for the Southeastern U.S. domain. This entailed the use of gridded projection factors developed by Prestemon and Mercer at USFS from the population and economics growth database provided by Woods & Poole Economics, Inc. for future years; base year fire activity statistics were obtained from the National Fire and Aviation Management Web Applications database (see http://fam.nwcg.gov/fam-web/ ). These data, however, showed severe underestimates of areas burned in the Southeast. Thus, the analysis was repeated with 2002 burned areas compiled for the Western Regional Air Partnership by Air Sciences, Inc., from which future-year annual burned areas were estimated. Modifications and data linkages to the Fire Scenario Builder (FSB), a stochastic model for distributing these estimates into daily areas burned (McKenzie et al., 2006), over the gridded domain, were also undertaken in the third and fourth year of the project.
Air Quality Modeling
Under a previous STAR grant, UNC investigators developed an air quality modeling system with integrated meteorology and chemistry (METCHEM) to fully couple regional-scale atmospheric dynamics with the chemistry-transport of trace chemical species. It is based on further development and refinement of three established models: the MM5 mesoscale meteorology model (Grell et al., 1994); the Multiscale Air Quality Simulation Platform (MAQSIP) (Mathur et al., 2005); and the SMOKE modeling system (Houyoux and Vukovich, 1999). 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 due to the presence of aerosols, resulting in a reduction in the shortwave radiation reaching the ground, and consequently, in the surface temperature and boundary layer height (see www.ie.unc.edu/cempd/projects/integrated ).
Figure 3. METCHEM simulation domains; domain D02 includes the impacts of PnET-predicted biomass changes on the fire emissions inputs.
In Year 1 of the project, a problem in the ISORROPIA thermodynamic model that determines the inorganic aerosol composition was identified and corrected. At high relative humidity, the model can lead to unrealistically high aerosol liquid water content (LWC). In consultation with the developer at Georgia Institute of Technology, the upper limit on relative humidity was reduced to 95 percent, and produced more realistic wet aerosol mass concentrations. To determine the aerosol radiative impact on meteorological variables, e.g., temperature profiles and surface solar heating, the METCHEM radiative transfer algorithm was updated in the indices of refraction for the aerosol constituents. In Year 2, this algorithm was further refined as a modular and numerically efficient parameterization of the Mie aerosol extinction calculation, and was found to reproduce the detailed Mie calculation to very good accuracy over a wide range of size parameter values. In Years 3 and 4, METCHEM was applied over the CONUS domain for 2-week test periods in the summer and winter of 2002. Results were compared against the IMPROVE monitoring network and the EPA Speciation Trends Network (STN) data. The model overpredicted ammonium and nitrate, and work began on diagnosing these overpredictions through a comparison of the RPO inventory data used in these simulations against the EPA NEI 2002, and through diagnostic analysis of the meteorological model configuration.
Conclusions:
The research showed the following: (a) future year fuel loads can be estimated at least for canopy fuel from the biomass predicted with a forest process model; monthly-average biomass estimates can be made from such a model for a regional domain such as the eastern United States. The difficulty of estimating input parameters makes it unfeasible to get a finer time resolution than monthly, but given the uncertainties inherent in estimating fuel loads from live biomass, this is a reasonable approach; (b) Estimates of future year areas burned are limited by the availability of reliable economic and demographic data, as well as projection/growth information for most emission sectors beyond 30 years at the selected spatial scale of 36-km or less. (c) The statistical model showed that areas burned by lightning rose significantly in 2020 and 2030, and decreased slightly in 2050 (to 31% higher than 2002 levels). It showed a steady decrease in domain-total human-caused fires in future years, down 30 percent from 2002 levels by 2050. The net effect was a decrease in the Southeast by 22 percent from 2002 levels for all fires. While climate change and the consequent increases in annual minimum temperature in parts of the domain (TX, LA) could cause the increases in areas burned by lightning, population and income growth, with increased break-up of landscape and better fire prevention methods appear to mitigate them, at least in the assumed growth scenario (A1) in the Southeast. Overall, the integrated modeling approach is valuable albeit too complex to address in all its aspects within a 3.5-year performance period.References:
Aber, J. D., S. V. Ollinger, and C. T. Driscoll, 1997: Modeling nitrogen deposition in forest ecosystems in response to land use and atmospheric deposition, Ecol. Modell., 101, 61-78. http://www.pnet.sr.unh.edu/onlinepubs/EcoMod-v101-p61.html
Grell, A. G., J. Dudhia, and , D.R. Stauffer, 1994: A description of the fifth-generation Penn State/NCAR Mesoscale Model (MM5). NCAR Technical Note NCAR/TN-398+STR, National Center of Atmospheric Research, Boulder, CO.
Houyoux, M., R. and J. M. Vukovich, 1999: Updates to the Sparse Matrix Operator Kernel Emissions (SMOKE) Modeling System and Integration with Models-3. Presented at The Emission Inventory Conference: Regional Strategies for the Future, Air & Waste Management Association, 26-28 October, Raleigh, NC.
Mathur, R., U. Shankar, A. F. Hanna, M. T. Odman et al., 2005: Multiscale Air Quality Simulation Platform (MAQSIP): Initial applications and performance for tropospheric ozone and particulate matter, J. Geophys. Res., 110, D13308, doi:10.1029/2004JD004918.
McKenzie, D.M., S.M. O’Neill, N.K. Larkin, and R.A. Norheim, 2006: Integrating models to predict regional haze from wildland fire, Ecol. Modell., 199, doi:10.1016/j.ecolmodel.2006.05.029.
Mercer, D. E., and J. P. Prestemon, 2005: Comparing production function models for wildfire risk analysis in the wildland–urban interface, Forest Policy Economics, 7, 782-795.
Sestak, M., S. O’Neill, S. Ferguson, J. Ching, and D. Fox, 2002: 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 CMAS Workshop, October 21-23, 2002, Research Triangle Park, NC. http://www.cmascenter.org/conference/2002/session5/fox_abstract.pdf
Journal Articles:
No journal articles submitted with this report: View all 21 publications for this projectSupplemental Keywords:
RFA, Scientific Discipline, Air, Ecosystem Protection/Environmental Exposure & Risk, climate change, Air Pollution Effects, Chemistry, 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:
The project website is http://www.ie.unc.edu/cempd/projects/FIRE/ .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.