NASA Logo

NTRS

NTRS - NASA Technical Reports Server

Back to Results
Investigating Uncertainty in Predicting Carbon Dynamics in North American Biomes: Putting Support-Effect Bias in PerspectiveA fundamental strategy in NASA's Earth Observing System's (EOS) monitoring of vegetation and its contribution to the global carbon cycle is to rely on deterministic, process-based ecosystem models to make predictions of carbon flux over large regions. These models are parameterized (that is, the input variables are derived) using remotely sensed images such as those from the Moderate Resolution Imaging Spectroradiometer (MODIS), ground measurements and interpolated maps. Since early applications of these models, investigators have noted that results depend partly on the spatial support of the input variables. In general, the larger the support of the input data, the greater the chance that the effects of important components of the ecosystem will be averaged out. A review of previous work shows that using large supports can cause either positive or negative bias in carbon flux predictions. To put the magnitude and direction of these biases in perspective, we must quantify the range of uncertainty on our best measurements of carbon-related variables made on equivalent areas. In other words, support-effect bias should be placed in the context of prediction uncertainty from other sources. If the range of uncertainty at the smallest support is less than the support-effect bias, more research emphasis should probably be placed on support sizes that are intermediate between those of field measurements and MODIS. If the uncertainty range at the smallest support is larger than the support-effect bias, the accuracy of MODIS-based predictions will be difficult to quantify and more emphasis should be placed on field-scale characterization and sampling. This talk will describe methods to address these issues using a field measurement campaign in North America and "upscaling" using geostatistical estimation and simulation.
Document ID
20010106499
Acquisition Source
Ames Research Center
Document Type
Preprint (Draft being sent to journal)
Authors
Dungan, Jennifer L.
(NASA Ames Research Center Moffett Field, CA United States)
Brass, Jim
Date Acquired
August 20, 2013
Publication Date
May 14, 2001
Subject Category
Environment Pollution
Meeting Information
Meeting: Uncertainty in Remote Sensing and GIS
Location: Southampton
Country: United Kingdom
Start Date: July 3, 2001
End Date: July 4, 2001
Sponsors: Remote Sensing and Photogrammetry Society
Funding Number(s)
PROJECT: RTOP 622-93-42-10
Distribution Limits
Public
Copyright
Work of the US Gov. Public Use Permitted.

Available Downloads

There are no available downloads for this record.
No Preview Available