Www.mtri.org A 2012 NASA-CMS Phase 2 study Lead Investigators: Nancy HF French, MTRI Don McKenzie, US Forest Service, PNW, FERA Eric Kasischke, University.

www.mtri.org

A 2012 NASA-CMS Phase 2 study

Lead Investigators:

Nancy HF French, MTRIDon McKenzie, US Forest Service, PNW, FERAEric Kasischke, University of Maryland

Presented by:

Michael Billmire, MTRI (WFEIS technical lead)

Regional Fire Emissions Products Regional Fire Emissions Products using the using the Wildland Fire Emissions Information SystemWildland Fire Emissions Information System

wfeis.mtri.org

Modeling Wildland Fire EmissionsModeling Wildland Fire Emissions

Fuel moistureFuel moisture

Modeling Wildland Fire Emissions: Modeling Wildland Fire Emissions: WFEISWFEIS

Fuel moistureFuel moisture

Burned area products provide:•burn area•location -> tells us fuel loadings, fuel moisture•day -> tells us fuel moisture

Options:•MODIS MCD64A1•MTBS burn perimeters•Daily progression (MODIS Active Fire interpolated to MTBS)•SmartFire 2011 NEI•Agricultural NEI (McCarty)

Modeling Wildland Fire Emissions: Modeling Wildland Fire Emissions: WFEISWFEIS

Fuel moistureFuel moisture

Fuel type (from 1-km FCCS map) linked to stratified fuel loadings (biomass)

Modeling Wildland Fire Emissions: Modeling Wildland Fire Emissions: WFEISWFEIS

Fuel moistureFuel moisture

Daily gridded weather •Daily fuel moisture interpolated from RAWS data from 1986 through 2013 at 40-km scale for all of CONUS

Modeling Wildland Fire Emissions: Modeling Wildland Fire Emissions: WFEISWFEIS

Fuel moistureFuel moisture

Consumption & Consumption & Emissions model Emissions model

(CONSUME)(CONSUME)

Consumption & Consumption & Emissions model Emissions model

(CONSUME)(CONSUME)

Modeling Wildland Fire Emissions: Modeling Wildland Fire Emissions: WFEISWFEIS

Fuel moistureFuel moisture

Main Project ObjectivesMain Project Objectives

• Improve WFEIS• Improve inputs: fuel loading (biomass), fuel moisture

• Develop methodology for uncertainty measurement for emissions estimations

• Produce regional emissions estimates

• Compare these estimates to CASA GFED estimates

Objective Objective

Improve WFEIS inputsImprove WFEIS inputs

aggregated classes

Biomass varies across instances of each class

Refining the fuels map for CONUSRefining the fuels map for CONUS(Co-I McKenzie)(Co-I McKenzie)

Each instance of a class (cell) is assigned a unique set of fuel loadings.

Refining the fuels map for CONUSRefining the fuels map for CONUS(Co-I McKenzie)(Co-I McKenzie)

We can use remote sensing products to help to more appropriately characterize canopy fuels and thus provide fuel load variation within a single fuel type

Tree cover

Non-tree cover

Bare ground

continuous values

14

Canopy fuel loadings in forested fuelbeds are generally lower. Exceptions in northern Rocky Mts, Northeast, Gulf coast, upper Michigan.

As expected, spatial variation is more “realistic” in MODIS-enhanced fuel layer.

Refining the fuels map for CONUSRefining the fuels map for CONUS(Co-I McKenzie)(Co-I McKenzie)

Result:The new 1-km MODIS-enhanced FCCS

15

Task: Improved Fuel Moisture InputsTask: Improved Fuel Moisture Inputs

Before: -by Level II CEC Ecoregion

Now: -by 40 km grid cell (interpolated from RAWS)

“Improved” = •better spatial resolution•updated through 2013

Develop methodology for assessing Develop methodology for assessing uncertaintyuncertainty

Achievements: – Developed a framework for estimating the various

components of WFEIS model uncertainty– Mathematically rigorous documentation of the (~60)

Consume regression equations– Identified all possible sources of error:

• Coefficient error for each regression-based prediction equation • Input measurement error (e.g. measurements from field work)• Classification error (e.g. accuracy of spatial data)

Provide Assessment of Uncertainty Provide Assessment of Uncertainty in Emissions Estimatesin Emissions Estimates

Achievements:

– Conducted case studies on two of Consume’s regression-based consumption equations

– The two studies were done for the shrub and for the western woody fuel strata

– These examples demonstrated both sample-based and Monte Carlo approaches to uncertainty and sensitivity analysis.

ObjectiveObjective

Produce regional emissions estimatesProduce regional emissions estimates

WFEIS Emissions EstimatesWFEIS Emissions Estimates

WFEIS Emissions EstimatesWFEIS Emissions Estimates

All available at:

http://wfeis.mtri.org/media/model_outputs

Region 11 – Mediterranean CaliforniaThis is the only ecoregion in the continental US with a Mediterranean climate – summers are hot and dry, and winters are mild. Droughts are common, with precipitation averaging from 200-1,000 mm per year. With irrigation, these features create a prime environment for high value agriculture. Native vegetation is dominated by shrubs, with patchy areas of grasslands and forests of evergreen and deciduous trees.

11.1 Mediterranean California

ObjectiveObjective

Compare to CASA GFED estimatesCompare to CASA GFED estimates

GFED comparison to WFEIS GFED comparison to WFEIS (with collaborator Collatz)(with collaborator Collatz)

Comparison between GFEDv3 (published emissions) and WFEISv0.3 for TENA/CONUS 2001-2010

GFEDv3 is the fire emissions module in CASA-GFED, one of the models used in the CMS-Flux pilot for terrestrial C flux estimates

GFED comparison to WFEIS GFED comparison to WFEIS (with collaborator Collatz)(with collaborator Collatz)

Summary of differences between major model components

GFED comparison to WFEIS GFED comparison to WFEIS (with collaborator Collatz)(with collaborator Collatz)

Modeled emission results comparison– Carbon: GFED ~50% of WFEIS

GFED comparison to WFEIS GFED comparison to WFEIS (with collaborator Collatz)(with collaborator Collatz)

Modeled emission results comparison– CO2: GFED ~95% of WFEIS

Project is complete, BUT we are currently seeking funding for the continued development/maintenance of WFEIS and are eager to get to work on the following important tasks:

– Full uncertainty characterization of the CONSUME model• Started under this project; statistical framework established and

demonstrated test cases for several of the ~60 consumption equations

• Once complete, the end goal is for WFEIS outputs (and any other modeling effort making use of the CONSUME model) to deliver uncertainty estimates

The Road AheadThe Road Ahead

Project is complete, BUT we are currently seeking funding for the continued development/maintenance of WFEIS and would love to get to work on the following important tasks:

• Integrate information on land disturbance from remote sensing and other geospatial products to update dynamically the representation of fuel loadings in WFEIS

The Road AheadThe Road Ahead

Project is complete, BUT we are currently seeking funding for the continued development/maintenance of WFEIS and would love to get to work on the following important tasks:

• Develop international versions of WFEIS

The Road AheadThe Road Ahead

Project TeamProject Team

Michigan Tech Research InstituteMichigan Tech Research InstituteNancy HF French, PINancy HF French, PI

Michael Billmire, Arthur Endsley, Nicholas Molen, Jessica McCarty, Reid Sawtell, Amanda Grimm,

Amy Howes, Kim Mobley

USFS Fire and Environmental Research and ApplicationsUSFS Fire and Environmental Research and Applicationsand University of Washingtonand University of Washington

Don McKenzie, Co-IDon McKenzie, Co-IRoger Ottmar, Susan Prichard, Rob Norheim

Eric S Kasischke, Co-I, UMdEric S Kasischke, Co-I, UMd

Jim Collatz, Collaborator, GSFCJim Collatz, Collaborator, GSFC

GFED comparison to WFEIS GFED comparison to WFEIS (with collaborator Collatz)(with collaborator Collatz)

Burned area comparison– GFED 2% higher than WFEIS

GFED comparison to WFEIS GFED comparison to WFEIS (with collaborator Collatz)(with collaborator Collatz)

Fuel available comparison– GFED 55% lower than WFEIS

GFED comparison to WFEIS GFED comparison to WFEIS (with collaborator Collatz)(with collaborator Collatz)

Combustion completeness comparison– GFED 15% higher than WFEIS

GFED comparison to WFEIS GFED comparison to WFEIS (with collaborator Collatz)(with collaborator Collatz)

Emission factors comparison– Carbon: GFED 5% lower than WFEIS

– CO2: GFED 80% higher than WFEIS