B51I-06
The 2014 Tanana Inventory Pilot: A USFSNASA partnership to leverage advanced remote sensing technologies for forest inventory
Friday, 18 December 2015: 09:15
2004 (Moscone West)
Hans-Erik Andersen1, Chad Ryan Babcock2, Bruce Cook3, Douglas C Morton3, Robert Pattison4 and Andrew Finley5, (1)USDA Forest Service Pacific Northwest Research Station, Seattle, WA, United States, (2)University of Washington Seattle Campus, School of Environmental and Forest Sciences, Seattle, WA, United States, (3)NASA Goddard Space Flight Center, Greenbelt, MD, United States, (4)USDA Forest Service, Pacific Northwest Research Station, Anchorage, AK, United States, (5)Michigan State University, Department of Forestry and Geography, East Lansing, MI, United States
Abstract:
Interior Alaska (approx. 50 million forested hectacres in size) is the last remaining forested area in the United States (US) where the Forest Inventory and Analysis (FIA) program is not currently implemented. A joint NASA-FIA inventory pilot project was carried out in 2014 to evaluate the utility of state-of-the-art high-resolution remote sensing information (lidar, hyperspectral and thermal airborne imaging) to support a future FIA inventory program in interior Alaska.
FIA plots were established at a 1:4 intensity (or 1 plot per 9,715 hectares) on a regular (i.e. systematic) hexagonal grid across the Tanana Valley State Forest and Tetlin National Wildlife Refuge; both of which fall within the Tanana valley of interior Alaska. The relatively sparse FIA field plot sample collection was augmented with samples of airborne remotely sensed data acquired with Goddard’s Lidar Hyperspectral and Thermal (GLiHT) imager to increase the precision of inventory parameter estimates. G-LiHT is a portable, airborne imaging system, developed at NASA Goddard Space Flight Center, that simultaneously maps the composition, structure, and function of terrestrial ecosystems. G-LiHT data supports local-scale mapping and regional-scale sampling of plant biomass, photosynthesis, and disturbance. The data is accurately georeferenced and can be matched precisely with field plot data that are georeferenced using survey-grade GPS. G-LiHT data was acquired in July-August, 2014 along single swaths (250 meters wide) spaced 9.3 km apart over the entire Tanana inventory unit (135,000 km2).
We examine three methodological approaches to estimate forest inventory variables of interest; focusing initially on aboveground biomass (AGB) estimation. The three estimation procedures include 1) the standard, fully design-based approach currently used by the FIA; 2) A model-assisted technique; and 3) a Bayesian multi-level modeling approach where the sampling design can be explicitly accommodated within the modeling framework. We plan to assess the accuracy and bias of the three approaches experimentally via simulation and using the field and remote sensing data collected during the 2014 Tanana Inventory Pilot project.
