Climet FireBGCv2 Simulation Design

¹USDA Forest Service, Rocky Mountain Research Station, Fire Sciences Laboratory

Abstract

The FireBGCv2 simulations for completion of the CLIMET project are structured to assess the differences between weather and fire regimes on characteristics of six watersheds (i.e., landscapes) across three National Parks in the Pacific Northwest, USA. This study will use a full factorial experimental design where climate and fire are the factors and there are four climate treatments or scenarios and three fire regime treatments. The four climate scenarios are constructed from modifications of the daily weather stream collected for each watershed using parameters provided by Phillip Mote, University of Washington. The scenarios are:

1. Historical climate (HC): The current weather stream without C02 ramp
2. Current climate (CC): The weather stream as it currently exists
3. Climate change alternative A (A2): Hadley A2 scenario (HadCM3 A2)
4. Climate change alternative B (B2): PCM B2 Mote scenario

The three fire regime scenarios attempt to span the range of fire management strategies for the six watersheds. They are:

1. Native Fire Scenario (NF): This fire regime is achieved by allowing the climate and vegetation to dictate the frequency and severity of fire. No suppression or fuel treatment activities are implemented.
2. Full Suppression scenario (SF): This fire regime only allows 10% of the ignited fires to burn. Once started, the fires can obtain any fire size. Severity is computed at the stand level
3. Wildland Fire Use scenario (WF): This fire regime allows 50% of the ignited fires to burn. Again, the size and severity of the fire are dictated by the vegetation and fuels.

Study Plan

The 12 simulation scenarios will be run on each of the six watersheds that comprise a east-west transect across the PNW. The six landscapes are:

1. Glacier National Park
   
   1. McDonald Drainage
   2. St. Marys Drainage
2. North Cascades National Park
   1. Stehekin Drainage
   2. Thunder Drainage
3. Olympic National Park
   1. Hoh Drainage
   2. Dungeness Drainage

1. Simulate landscapes for 1,000 years
2. Ramp climate from current to climate scenario over 100 years
3. Report simulated landscape characteristics every 10 years for the first 100 years
4. Report simulated landscape characteristics every 50 years for the remaining years
5. Generate landscape maps every 50 years
6. Use a 3 km buffer around each watershed to define the simulation landscape

A full suite of landscape characteristics will be generated to use in the full factorial ANOVA approach. These are the diagnostic metrics or variables that will be used to detect differences between climate and fire regime scenarios. These variables are computed for each output interval and the value at each output interval is considered a statistical observation. Here is the list of variables by major category:

1.  Landscape composition – these are the area occupied by the various categories that describe landscape composition.
   1. Area by cover type
   2. Area by structural stage
2. Landscape structure – these are the spatial metrics computed from the generated maps analyzed at the cover type level.
   1. Mean patch size
   2. Largest patch index
   3. Contagion
3. Fire – these are the spatial characteristics of the fires across the simulation time span.
   1. Area burned
   2. Number of fires
   3. Fireline intensity
   4. Flame length (spread rate?)
   5. Area in each fire severity class
   6. Smoke production
4. Wildlife habitat – The model predicts a habitat availability index for the entire landscape for each reporting interval. We will parameterize FireBGC to generate habitat availability values for three wildlife species.
   1. Grizzly bear habitat availability index
   2. Elk habitat availability index
   3. Lynx habitat availability index
5. Fuels – These are values that quantify various characteristics about the fuelbed that is critical for fire management.
   1. Area in each of the Anderson (1982) 13 fuel models
   2. Canopy bulk density
   3. Canopy base height
   4. Coarse woody debris loading
   5. Fine woody debris loading
   6. Snag density
6. Ecosystem processes – these are the values of various biogeochemical computations by the BGC engine in the center of FireBGC.
   1. Net primary productivity
   2. Net ecosystem production
   3. Evapotranspiration
   4. Standing Crop (total carbon loading)
   5. Basal area
   6. Stream discharge
      

Status

Here is the schedule that we’ve set for the completion of the CLIMET simulations.

• April-May 2005: Validate model results, revise model, add additional options to model. Run model to create equilibrated BGC parameters.
• June-December 2005: Run FireBGCv2 with the climate and fire scenarios. Generate output and evaluate for realism and consistency.
• January-April 2006: Analyze model results and write various papers.

This schedule assumes that we will have at least 6-10 processors available to perform the calculations over the June-Dec period.

Funding Organizations

This study has been funded by the JFSP to explore the effects of climate change and wildland fire use on other watersheds in the southwest – Rincon Wilderness, Gila Wilderness, Sequoia Kings Canyon NP, and Yosemite NP.

Products

The results generated from this simulation effort will be published in three or maybe four journal articles as follows:
Keane, RE. FireBGCv2 – Model documentation. USDA RMRS General Technical Report.

Keane, R.E., A. Keyser. Effects of climate change and fire management on landscape characteristics for six disparate watersheds in the Pacific Northwest, USA. Ecological Modelling.

Keyser, A., R.E. Keane. Effects of climate change and fire management on wildlife habitat. Journal of Wildlife Management.

Keane, R.E., A. Keyser. Effects of climate change on fire and fuel dynamics across the Pacific Northwest, USA. International Journal of Wildland Fire.


The publication of these two or three papers will hopefully successfully complete the FireBGC simulation phase of the CLIMET project. However, we plan to continue with the FireBGCv2 modeling of the CLIMET watersheds to explore a number of other modeling and landscape ecology areas. Listed below are a set of simulation studies that we have identified for implementation under the Landscape Assessment Team of the Fire Ecology and Fuels Research Project. They are:

1. Model Compartment Comparison (Keane). An investigation of the sensitivity of simulating landscape processes using a variety of simulation approaches. We will compare the gap and BGC simulations of the FireBGCv2 model with the simple LANDSUMv4 model outputs to identifiy significant gains or losses in prediction results as a consequence of simulation resolution. We will also using three fire spread models of increasing complexity to identify similar simulation advantages and disadvantages.
2.  Wildfire vs RxFire Tradeoff (Keane, Keyser). We will simulate FireBGCv2 under 10 fire management scenarios where wildfire and prescribed fire levels are set and the effect of these levels are compared against landscape characteristics. A proposal for this study was submitted to the Joint Fire Sciences Program in 2002 but was NOT funded.
3. Implementing Climate Change Into Management Plans (Keane, Keyser). Here, we would craft a set of simulation scenarios that create a table that showed the effect of climate change on a variety of landscape characteristics. This table can then be used by management to adjust their estimates of ecological consequences based on trends in the table. A proposal for this study was submitted to the Joint Fire Sciences Program in 2003 but was NOT funded.
4. Determining Climatic Thresholds (Keane, Keyser). This study would attempt to use FireBGC as an iterative tool to identify climatic thresholds for a suite of landscape characteristics where these characteristics drastically change once the threshold is reached.