www.mtri.org

Nancy HF French*, Don McKenzie, Tyler Erickson

*Presented at the 2012 ACCENT-IGAC-GEIA Conference, 11-13 June Toulouse, France

Wildland Fire Emissions Information System for retrospective fire emissions mapping for North America

Fire & Carbon

GFED: The global amount of carbon combusted through fires is about 2.0 Pg C year-1, or about 22% of global fossil fuel emissions. van der Werf et al. 2010. Atmos. Chem. Phys. Discuss.,10:16153-16230, www.atmos-chem-phys-discuss.net/10/16153/2010/

CBM-CFS3: From 1990 to 2006 Canada’s managed forest was a Carbon Sink on average, but it was a Carbon Source in years with large burned area. Kurz, W.A. et al. 2008. PNAS 105:1551-1555.

Fire Emissions Project

Purpose: Improve access to fire emissions model inputs and results for targeted users Provide best estimates of total carbon emissions and some emission components for retrospective fire emissions mapping – Geospatially at 1km resolution – At daily to annual temporal resolution

Progress: Improved data & model for retrospective fire emissions estimation Development of a prototype web-based system (WFEIS) for US Regional-scale retrospective estimates of US fire emissions based on two burn area data sets

- Supported by a 3-year grant from NASA Carbon Cycle Science Program & Applied Sciences Program

- Additional funding under another NASA Applied Sciences - Actively seeking additional funding to continue development

4

WFEIS Project Team

Nancy French, Ben Koziol, Tyler Erickson, Mike Billmire, Jessica McCarty

Don McKenzie Roger Ottmar Ernesto Alvarado

Bill de Groot

Eric Kasischke

Research Associates & Interns: Liza Jenkins, Eric Keefauver, Arthur Endsley, Reid Sawtell, Susan Pritchard, Ron Kempker, Jef Cieslinski, Marlene Tyner, Christina Nolte, Peter Gamberg, Naomi Hamermesh

Pacific Northwest Research Station, USDA Forest Service

Department of Geography, University of Maryland

Great Lakes Forestry Centre, Canadian Forest Service

Fire Emissions Modeling

Non-spatial, plot-level (Consume 3.0, FOFEM 5.7) Semi- or Pseudo-spatial (Canadian FBP, CanFIRE) Global, coarse-scale (GFED) Active fire estimates of burn area (FINN) Many other efforts in place for emissions inventory/air quality applications (e.g. US-EPA)

WFEIS uses moderate resolution data of burn area and biomass to make regional-scale fire emissions estimates

-- CONSUME 3.0 (USFS)

-- FOFEM 5.7 (USFS)

-- CanFIRE -- Canadian FBP -- GFED (NASA) -- FINN (NOAA)

Other fire emissions models:

Combustion Factors - ß

Area Burned - A

Fuel Consumption

Emission Factors - EF

Emissions

Fuel Loading - B

WFEIS Estimation of Fire Emissions

Daily weather ĺ�)XHO�PRLVWXUH

Location and day of fire

Fuel (vegetation) type (FCCS)

Fuel load (biomass) and fire behavior

Consumption & Emissions model

(Consume)

Output: Spatial representation of emissions

Burn Area Datasets

Perimeters from Monitoring Trends in Burn Severity (MTBS) http://mtbs.gov

MODIS-derived Direct Broadcast Burn Area Product (MCD64A1) see Giglio et. al 2009 – 500 m spatial resolution – Burn cells tagged by

approximate burn date – North America-wide

for 2001 to present

7

8

Fuel Consumption and Emissions: CONSUME

CONSUME estimates fuel consumption and emissions for prescribed and wildland fire. It imports fuelbed data directly from the FCCS, and can be used for all forest, shrub, and grassland types in North America.

Litter

“Flash” Fuels 1-hr

Log 100-hr

Pyrolysis zone Flame-Available Fuel

Smolder-available Residual-available

Estimates combustible biomass of woody fuels in each of the three stages of combustion.

Predicts fuel consumption, pollutant emissions, and heat release based on:

• fuel loadings • fuel moisture • and other environmental

factors

http://www.fs.fed.us/pnw/fera/research/smoke/consume/

Low-intensity prescribed fire and high-intensity crown fire consume different proportions of each stratum in each combustion phase.

9

Fuel Loading: Fuel Characteristic Classification System (FCCS)

FCCS provides an comprehensive description of fuel layers.

Compiled from scientific literature, fuels photo series, fuels inventories, and expert opinion, and represent fuel conditions at multiple scales, from single plots to 1-km cells or larger.

Fuelbed Map • Fuelbeds are mapped via crosswalks to satellite-derived vegetation and land cover, at scales from < 25 m (landscape applications) to >36 km (continental and global applications).

• USGS Landfire project has developed a crosswalk from 30-m scale existing vegetation maps

• For WFEIS, forest & rangeland fuels are mapped to 1-km

http://www.fs.fed.us/pnw/fera/fccs/index.shtml

includes fuel loadings by type

Fuels Mapping: Mapping standard FCCS fuelbeds

30-m resolution

1-km resolution

234 fuelbeds 5 lost in aggregation

Problem: – Original FCCS map does not include all land types with vegetation fuels; – Missing types primarily are areas that are urban, bare, or agricultural; – When fires occur in these area, CONSUME calculates their emissions as 0. – Fires do often occur in agricultural areas.

Solution: – The Cropland Data Layer (CDL) contains

detailed information about agricultural fuelbeds in many areas within the FCCS gaps.

– Using ERDAS Imagine Model Maker, we merge the CDL with the FCCS to create an integrated fuelbed map.

• Create new agricultural fuelbeds, reclassify CDL merge giving priority to FCCS in areas where the CDL is vague (e.g. Deciduous Forest) and giving priority to the CDL in all other locations.

Fuels Mapping: Filling in the Gaps with the Cropland Data Layer

FCCS (30m resolution)

Agricultural areas missing data

CDL (30m resolution) Merged FCCS/CDL Layer

(30m resolution)

Fuels Mapping: Filling in the Gaps with the Cropland Data Layer

Merged FCCS/CDL layer (30m resolution)

• 242 Total fuelbeds • 212 original FCCS fuelbeds • 15% of total CONUS area ‘filled in’ with croplands • New fuelbeds from CDL publically available; fuel loadings

vary by crop type

15

WFEIS Components

16

CONSUME

FCCS Fuelbed

Fire Extent & Timing

Daily Weather/

Fuel Moisture

KML

SHP

TXT Shrub

Consumption Canopy

Consumption

User Input System Output

NetCDF

Select By: • AOI/lat,long • Fire Name • Place Name • Time Select Fire Perimeter: • Landsat (MTBS) • MODIS (MCD64A1)

GeoTIFF

Modify Pre-set Inputs

(optional) WFEIS Back End

Geospatial Emissions Modeling

WFEIS Web-accessible Framework

The WFEIS website allows for two approaches for making fuel consumption and emissions estimates

Method #1: WFEIS Emissions Calculator: http://wfeis.mtri.org/ Method #2: WFEIS responds to queries submitted via properly encoded URL requests: it implements a RESTful Web API.

Django Web Framework

Spatial Database

Backend

RESTful Interface

User Interface

Python Scripts

HTTP Access

Web Browser Access

CONSUME

WFEIS Framework

URL-formatted QUERY

http://wfeis.mtri.org/api/emissions/fuelbed=fccs1km/burnedarea=mtbs/ecoregion=western/1000hr_FM=None/Duff_FM=None/CanopyPerConsume=None/PercentBlack=50/combustion_stage=total/stratum=total/output_units=tonnes/emistype=carbon/map.kml?DRNG=2002-07-01,2002-07-31&ROI=StateProvince,usa-OR

WFEIS Output

WFEIS Output Formats

z KML – Google Earth format

z ESRI Shapefile – suitable for desktop GIS

z NetCDF – suitable for atmospheric scientists

z GeoTIFF – a georeferenced image

z Text Report – a summary report

Aggregating Output

WFEIS can be used to consider annual fire emissions across the Contiguous US

• “batch” API queries allow for computing a variety of scenarios • This example shows results using the MODIS MCD64A1 burned

area product CONUS total area is approx 8,000,000 km2

Year Burned area (km2)

Total consumption (Tg)

Consumption normalized (kg/m2)

Total CO2 (Tg)

2001 5462 9.18 1.68 14.49 2002 12885 45.49 3.53 71.84 2003 11886 43.20 3.63 68.70 2004 5566 11.32 2.03 18.38 2005 13428 16.51 1.23 26.73 2006 23779 43.64 1.84 69.25 2007 25964 68.68 2.65 109.30 2008 7749 37.32 4.82 58.48 2009 14371 19.91 1.39 32.27

Aggregating Output

Consumption & emissions for ecoregions of the Contiguous US • example is for 2003

CONUS Ecoregion

Burned area (km2)

Total fuel consumption

(Tg)

Fuel consumption normalized

(kg/m2)

Total CO2 (Tg)

Northwestern Forested Mountains 3853 26.74 6.94 41.98

Mediterranean California 3085 8.29 2.69 13.58 Great Plains 1451 3.54 0.26-10.19 5.73

North American Deserts 976 0.44 0.43-0.45 0.71 Tropical Wet Forests 920 0.05 0.05 0.08

Temperate Sierras 696 1.50 2.15 2.40 Eastern Temperate

Forests 418 0.80 0.22-2.21 1.33

Southern Semi-Arid Highlands 367 0.54 1.46 0.88

Northern Forests 71 1.06 15.01 1.63 Marine West Coast

Forest 50 0.25 4.91 0.40

Comparison of Models Five fire events:

-- 2002 Biscuit fire in southern Oregon -- 2003 Montreal Lake central Saskatchewan -- 2004 Boundary fire in interior Alaska -- San Diego County, California 2003 and 2007

Results: -- Models generally agree (within 25% of each other) but vary due to model assumptions

-- Vegetation fuel density, structure, and condition (fuel moisture) are important drivers of emissions variability

-- Global-scale GFED modeled emissions are consistent with landscape/regional-scale estimates

0.00

0.50

1.00

1.50

2.00

2.50

3.00

3.50

10.15 (SanDiego

07)

17.76 (SanDiego

03)

28.55 (Montreal

Lake)

33.89 (Boundary)

89.97 (Biscuit)

Area

Nor

mal

ized

Carb

on E

miss

ions

(kg

m-2

)

Fuel Loading (kg m-2)

French, N.H.F. et al. (2011), "Model comparisons for estimating carbon emissions from North American wildland fire," Journal of Geophysical Research, 116, G00K05.

Effect of Burned Area on Carbon Emissions

For the Biscuit fire case, the WFEIS model was run with two burn area maps. The burn area and total carbon emitted is higher with the Landsat-derived map by about 16% Area normalized emissions were very similar, but vary due to use of daily weather information and the type of vegetation in the burn

Landsat: 200,400 ha 3.10 kgC/m2 6.20 TgC MODIS: 170,000 ha 3.07 kgC/m2 5.22 TgC

Data Access

http://wfeis.mtri.org

Fuels maps available from USFS FERA lab http://www.fs.fed.us/pnw/fera/fccs/index.shtml

Accessible soon from the NASA-DAAC at ORNL:

http://daac.ornl.gov/

Questions?