Project Methods and Materials
Objective 1. Measurement and Modeling of Down Woody Debris and Fuels
We proposed to establish a permanent plot network on the Alligator River and Pea Island National Wildlife Refuges and the Air Force Dare County Bombing Range modeled on USDA Forest Service FIA P2 and P3 plots to measure and characterize live biomass and pre- and post-burn down deadwood (DWD) on two prescribed fire sites and two control plots in each of the first two years of the study. We will use field protocols based in methods establish by the USDA Forest Service in Field Instructions for Southern Forest Inventory (http://fia.fs.fed.us/library.htm#Manuals). The collection of DWD data will use a line-intersect method to sample down wood along linear transects. Plot-level data on the amount, distribution, and characterization of DWD can be related to the detailed attribute data for other ecosystem components on the same plot (i.e., shrub and herbaceous understory, standing dead, and live biomass) (Figure 1.).
| Figure 1. |
|
| Click to see larger image and plot description |
|
| Taking field data for FIA plots. |
We will acquire USGS NAP color infrared (CIR) negatives of stereo aerial photography for four study areas in the first and second years of the study and collect field data on fuel loads and/or fuel accumulation. Aerial photos will be scanned to generate digital coverages and stereo models for interpretation as well as orthorectified mosaics of the study areas. We will incorporate existing GIS vegetation data from the USFWS and the US Air Force, or classify newly acquired aerial photography (US Air Force FY'2004 leaf off 1:600 scale color infrared) using onscreen stereoscopic techniques to create a digital vegetation database. Fire fuel data from field based sample plots, digital photos, and vegetation data will be used to develop fire fuel polygons. Additional field data will be used to assess the thematic accuracy of the vegetation classification, the positional accuracy of the digital orthophoto mosaic, and the fuel load polygons. Metadata will be created for the digital orthophoto mosaics and vegetation and fuel load databases.
 |
| Delineating vegetation alliances in ArcGIS using 3D color infrared photography |
Objective 2. Prescribed Burning Emissions Monitoring and Modeling
In order to assess emissions of fine PM from prescribed fire, we will investigate emissions from this burning practice during the course of prescribed burns for the fall and spring of 2005 and 2006. Mass balance techniques will be used to support flux measurements of dioxin, methyl chloride, methyl bromide, and other compounds. Grab samples (stainless steel canisters for trace gases, filter packs for PM, and polyurethane foam traps for semi-volatiles (Hays et al., 2002) will be collected in the plumes of the fire during both flaming and smoldering conditions. These samples will then subjected to particle and total gaseous carbon analysis using a thermo-gravimetric analysis and gas chromatography/mass spectroscopy (GC/MS). Total emissions will then be determined by multiplying emission ratio by the estimate of total fuel carbon consumption. Total fuel carbon by mass is approximately half of the fuel dry weight following the general cellulosic molecular formula of C6H9O4.
We will sample air that has cooled to approximately ambient temperature (within 100 meters of the fire) to allow partitioning of semi-volatiles between the gas and aerosol phase. We will then sample directly through PM2.5 cyclones using high volume pumps, onto 47 mm quartz (for organic PM components) and teflon filters (for inorganic ions and elements). These will be backed by polyurethane foam plugs for quantitative analysis of semi-volatile organic compounds that pass through or are volatilized from the filters. Trace gases (methane, C2-C12 volatile organic carbon (VOC )) will be characterized separately using Summa stainless steel canisters and GC/MS. Target compounds in the gas and PM phase will include saturated (alkane) and unsaturated hydrocarbons, aldehydes, ketones, organic acids, and polycyclic organic hydrocarbons (PAH). We will also measure CO via gas filter correlation (Teco 46C) and CO2 via infrared gas techniques (Licor 7000) to characterize carbon fluxes for the mass balance flux techniques and to characterize the nature of plume dispersion and proximity to the combustion. We will also use the CO/CO2/VOC measurements to help us in chemically identifying the flaming and smoldering stages of the fires.
|
| EPA sampling equipment during prescribed fire operations at the Croatan National Forest. |
We plan to quantitatively analyze these samples to determine emission factors for individual trace gases. These will be compared with factors from other fuel types. We also intend to compare current source apportionment chemical fingerprints from these fires with those from our laboratory and "burnhut" studies (Hayes et al., 2002). This will allow us to assess these signatures with in situ data. We will also compare emission factors and fluxes with any available data from wildfire emissions from corresponding forest types.
Objective 3. Smoke Modeling
The field validation of PB-Coastal Plain model will be led by Gary Achtemeier. PB-Coastal Plain is intended to extend the theory of PB-Piedmont to model for land/water interfaces and to include differences in vegetative land use. Because of the complexity of the meteorology of the Coastal Plain and the number of assumptions that must be made to develop a model that will run on a PC-environment in faster than real time, PB-Coastal Plain must be extensively validated with observations. Gaining the required observations is difficult for following reasons: (1) Coastal burn locations are more inaccessible than inland sites - there are fewer roads; (2) The meteorology of the Coastal Plain is highly complex - data from distant weather stations will not represent local weather; and (3) PB-Coastal Plain requires validation data not usually available through standard measurements.
|
| North Carolina Division of Forest Resources RAWS data collection unit positioned at the end of Point Peter Road on the Alligator River NWR. |
Data that will be collected for validating PB-Coastal Plain are: (1) Depth of the nocturnal boundary layer over the course of the night; (2) Temperature of the nocturnal boundary layer over the course of the night; (3) Ground level (2m) measurements of wind speeds down to 0.1 m/sec, temperature, and relative humidity from a dense network of surface weather stations surrounding the burn site; and (4) Observations of smoke location. PB-Coastal Plain is an operational numerical weather and smoke model designed to simulate near ground smoke transport at night. PB-Coastal Plain must be capable of modeling smoke on fine land use scales. The solution is a dual-grid nested model. The large grid with spacing 600-900 m will be used to simulate land/water interactions over a 200 km square surrounding the burn site. The small grid with spacing 30-150m will simulate smoke movement over variable land use.
PB-Coastal Plain will be validated for intended use in three ways. First, as a predictive model, it will be coupled with output from a mesoscale weather prediction model such as MM5 for use in pre-burn planning. Second, as a "nowcast" model, it will be coupled with hourly weather data to simulate where smoke is currently moving. Nowcast information will aid in decisions identifying smoke prone locations where smoke sentries need to concentrate their patrols. Third, as a diagnostic model, PB-Coastal Plain will be used as an educational tool for training land managers to anticipate smoke conditions over the lands they manage and surrounding landscapes under specified meteorological conditions.
The AirFIRE team, led by Dr. Sue Ferguson, will provide a forecast of expected weather and smoke behavior before each experiment, gather on-site meteorological information, and run and test the BlueSky smoke prediction system (www.fs.fed.us/bluesky). The forecast is necessary to help anticipate fire and smoke behavior and to determine the most effective observational configuration. On-site meteorological information will be used to validate and improve the weather components of BlueSky and PB-Coastal Plain. A standard configuration of BlueSky will be run to help with pre-burn forecasting and be used to demonstrate the uncertainty in predicting smoke in the region. An enhanced configuration of BlueSky will be run with measured information from each experiment to help quantify areas of needed improvement.
A forecast of expected weather and smoke behavior will be accomplished by expert analysis of all available climate information, topography, and inclement weather - including the high-resolution meteorological predictions produced by the Southern High Resolution Modeling Consortium ( SHRMC - http://shrmc.ggy.uga.edu/) and in collaboration with local fire weather forecasters. In addition, we will integrate SHRMC weather products, pre-burn fuel load, and anticipated ignition pattern into a standard configuration of the BlueSky system and analyze its predicted smoke behavior. The forecast will help us develop the integrative framework for applying BlueSky to the southeastern U.S., help test both SHRMC and BlueSky products in an operational setting, and prepare effectively for each experiment by anticipating weather and smoke conditions. Such forecasts are being produced regularly by the AirFIRE staff in support of prescribed burning by the Montana-Idaho Airshed Group and have aided over a dozen experimental burns in Florida, Georgia, Washington, Oregon, and Alaska.
The observations will include four surface weather stations at strategic points around the fire perimeter to capture surface drift smoke and influencing weather. Each station measures wind, temperature, humidity, and carbon monoxide every 5 minutes. An additional two stations are further outfitted with a particle sampler and are placed in expected downwind locations away from the fire perimeter to capture outflow rates. A tethersonde is deployed each night of the experiment to measure the vertical structure of the smoldering plume. The tethersonde measures wind, temperature, humidity, carbon monoxide, carbon dioxide, and particle concentrations and is launched hourly each night of burning through the depth of the inversion, sampling every 10 seconds. The location of each sampling station will be based on the topographic configuration of the area, expected weather, and anticipated fire and smoke behavior. The meteorological data will be monitored during each experiment to ensure quality control. At the end of each experiment, AirFIRE will further check the data and configure it into useful formats for the BlueSky system and PB-Coastal Plain model. This sampling methodology has been successfully deployed at the FROSTFIRE experimental burn in Alaska (Ferguson et al., 2003) and the surface measurements have been tested during experimental burns in Washington, Oregon, and the Piedmont region of northern Georgia with success.
Before each experiment we will run a standard configuration of the BlueSky smoke modeling system (www.fs.fed.us/bluesky), which uses the CALPUFF dispersion model. The system will be rerun following each experiment in an enhanced mode that adjusts available pre-burn information with information that was measured during the experiment. The two runs will be compared to quantify uncertainty in the modeling system and determine areas of needed improvement. The standard and enhanced BlueSky output will be made available to the PB-Coastal Plain development group to evaluate the degree of improvement afforded by its high-resolution computations.
Objective 3. Technology Transfer
The most important product of the proposed research is the development of computer based, real time decision support tools that can be easily operated by land managers for the implementation of smoke and fire management plans at the local level. It is also important that the tools developed for one location can be applied by land managers at other Coastal Plain sites in the southeastern US. As regards to the fire behavior and smoke models, the project will place into the hands of land managers in the North Carolina Coastal Plain, an operational version of the models. This will include the model software and an operations manual on how to run the models. Training will be provided for those interested in using the models as part of their prescribed burn decision-making process. The proposed study is a partnership between the research, the regulatory community, and on-the-ground fire managers at the wildlife refuge and military training range. All the data that will be collected is designed to meet the current information needs of land managers and all the decision support tools will be developed in partnership with land mangers during the entire three years of the study.