NASA Ames Develop Program / USDA / Fremont Winema Forest Carbon Management Project 2005
Project created by student team at NASA Ames Research Center DEVELOP Program 2004-2005.
Published on: Mar 3, 2016
Transcripts - NASA Ames Develop Program / USDA / Fremont Winema Forest Carbon Management Project 2005
Carbon Management in the Fremont-Winema National Forest, Oregon
Casey Cleve, San Francisco State University
Catherine Huybrechts, San Francisco State University
Kevin Hsu, Stanford University
Michael Jin, Stanford University
Mentors: Dr. Joseph Skiles, NASA Ames Research Center and Cindy Schmidt, San Jose State University
DEVELOP is a student-run, student-led internship program that extends NASA Earth Science research to local, county, state, and tribal governments and communities. In this project, students
at NASA Ames Research Center in California provided managers of the Fremont-Winema National Forest with data and analysis for carbon management. The goal of this project was to
demonstrate how different harvesting scenarios impact forest carbon budgets and to provide fire modeling outputs to assist managers in the mitigation of disturbances that could potentially
impact the forest’s carbon budget. The figures below display major components of this project, including: 1) a project flowchart, 2) model outputs from a NASA developed carbon simulation
model, CASA, and 3) fire behavior characteristic model outputs from FlamMap. DEVELOP is funded by the Applied Science Program at NASA Headquarters.
Study Area Vegetation Map
• The Fremont-Winema National Forest is comprised of over two million acres.
• Oregon is a significant area of study as it is the leading provider of lumber in the United States.
Project Flow Chart Dominant Vegetation
Species and Land cover
Fremont-Winema National Forest
• Within Fremont-Winema Forest, resource managers harvest 12 million ft3 of timber annually.
Field Data Satellite Imagery
In July 2004, the team spent a week collecting data and making observations in the forest Scenarios
and surrounding area. A forest service ecologist (left photo) and a geomorphologist (right Canopy Cover Basal Area
photo) assisted in vegetation species and land cover identification. The map was created using 4 2003 Landsat 5 images, which were mosaicked and georeferenced in
Climate Data preprocessing. 281 training sets, created from various sources, were input to a supervised classification and
reduced to the 22 class map you see here. Largest error was between selective cut, agriculture, and rangeland
classes. To mitigate this error, bands 3_4 and 4_5 feature space, and the spectral signatures of pure pixels
Weather/Fuel Moisture were used refine and separate classes.
GPS & Elevation
Forest Harvest Carbon Simulations
Slope/Aspect Anderson Fuel Type The team worked with NASA scientist Dr. Christopher Potter, the P.I. for the NASA-CASA model. The model
estimates Net Ecosystem Productivity (NEP) in grams of carbon based on climate, FPAR, soil texture, and
elevation. Two different carbon level scenarios - consisting of high and low slash values - were chosen for
Soil Data three variable climate regions throughout the forest. Simulations were conducted at three different 1m2 plot
Fire Behavior Characteristic Mapping Fire Behavior Model
locations, which were chosen based on climate variability (high to low precipitation). Re-growth estimates were
conducted using NASA-CASA to demonstrate the change in NEP after harvest.
FlamMap NASA-CASA Model Total Net Ecosystem Productivity Loss
Rate of Spread Flame Length
Rate of Spread
Simulated Fire Behavior Predicted Carbon Simulations
(meters/minute) Flame Length
None No data
Harvest location simulation 1
On the graph, the curve represents
NEP after harvesting a 1 m2 plot of
* distance between flame tip and midpoint of flame at flame base DEVELOP Customers timber, and subsequently 100 years
of simulated re-growth.
Fremont-Winema National Forest
The map on the left is Rate-of-Spread in meters per minute, showing how fast a fire moves across a landscape.
However, ROS by itself does not predict fire risk, since both a fast-spreading fire and a slow spreading fire can be hard
Harvest location simulation 1 (high precipitation) would take approximately 15 years for annual NEP to
to contain. We therefore created the flame length map on the right which is useful as it represents fire intensity. When Bureau of Land Management return to net-zero NEP. Harvest location simulation 2 (low precipitation) would take more than 21 years of
rate of spread (or speed) and flame length (or intensity) are combined it provides an estimate of fire risk. As these maps
growth for annual Net Primary Productivity (NPP) carbon to level with carbon losses from slash
were created to identify target areas for fuel-load reduction treatments, forest management may be interested in the
decomposition. Harvest location simulation 3 (low precipitation), had the same results as harvest location
south-east region enlarged between the maps, which under the conditions input to the FlamMap model would have an Klamath Tribe simulation 2 in that it would take more than 21 years of growth for NPP to level with carbon losses from
extremely fast moving fire with flame length extending approximately six-meters high. Fuel load reduction treatments,
slash decomposition. For both harvest locations 2 and 3, there is a net loss of 2100-3700g C, because NPP
such as clearing slash and fallen trees to prevent large fires, could act as a management tool in preserving the forest’s
slows down after 50 years of re-growth. All 3 harvest location simulations support the finding that NEP
carbon budget. Crater Lake National Park
carbon will never return to the level it was before harvesting.