An integrated, observation-based system to monitor aboveground forest carbon dynamics in Washington, Oregon, and California

Robert E Kennedy, Oregon State University, rkennedy@coas.oregonstate.edu (Presenter)
Matthew Gregory, Oregon State University, matt.gregory@oregonstate.edu
Janet L. Ohmann, Oregon State University, janet.ohmann@oregonstate.edu
Heather Roberts, Oregon State University, heather.roberts@oregonstate.edu
Neeti Neeti, Woods Hole Research Center, neeti.squared@gmail.com
David Miller, University of California, Santa Barbara, dmil1991@gmail.com
Zhiqiang Yang, Oregon State University, zhiqiang.yang@oregonstate.edu
Warren B. Cohen, USDA Forest Service, warren.cohen@oregonstate.edu
Van Kane, University of Washington, vkane@u.washington.edu
Jonathan Kane, University of Washington, jontkane@uw.edu
Scott L. Powell, Montana State University, spowell@montana.edu

Because carbon pools and fluxes on wooded landscapes are constrained by their type, age and health, understanding the causes and consequences of carbon change requires frequent observation of forest condition and of disturbance, mortality, and growth processes. As part of USDA and NASA funded efforts, we are building an empirical monitoring system that integrates time-series Landsat imagery, Forest Inventory and Analysis (FIA) plot data, small-footprint lidar data, and aerial photos to characterize key carbon dynamics in forested ecosystems of Washington, Oregon and California. Here we report yearly biomass estimates for every forested 30 by 30m pixel in the Oregon Cascade mountains from 1990 to 2010, including spatially explicit estimates of uncertainty in our yearly predictions. Total biomass at the ecoregion scale agrees well with estimates from FIA plot data alone, currently the only method for reliable monitoring in the forests of the region. Comparisons with estimates of biomass modeled from two small-footprint lidar acquisitions in overlapping portions of our study area show general patterns of agreement between the two types of estimation, but also reveal some disparities in spatial pattern potentially attributable to age and vegetation condition. Using machine-learning techniques based on both Landsat image time series and high resolution aerial photos, we then modeled the agent causing change in biomass for every change event in the region, and report the relative distribution of carbon loss attributable to natural disturbances (primarily fire and insect-related mortality) versus anthropogenic causes (forest management and development).

Presentation Type:  Poster

Session:  Theme 4: Human influence on global ecosystems   (Mon 4:30 PM)

Associated Project(s): 

• Kennedy, Robert: Integrating and Expanding a Regional Carbon Monitoring System into the NASA CMS ...details

Poster Location ID: 68