There is increasing and broad-based interest in the use, management and protection
of Minnesota forests. This interest creates extensive needs for improved information
to support policy and management decisions, some of which can be met with satellite
remote sensing. Remote sensing using aerial photography is an integral part of
forest inventory and management, but operational use of satellite data has been
limited. The goal of eForestl is to integrate satellite remote sensing into
forest inventory and management at all levels - local, state, federal, corporate
and private. The temporal frequency and potential timeliness of satellite data
products have a good deal to offer for certain field applications. We have pursued
the following three major objectives spanning the range of strategic inventory
to field level management:
- Improve the categorical resolution and accuracy of forest
classification and change detection by integrating the best combinations
of sensor data and analysis methods. We will systematically define/develop,
evaluate and demonstrate image processing protocols to convert these new
image data into meaningful information for forest inventory and management.
- Extend field data to landscapes by implementing a promising non-parametric technique
for estimation of forest variables from satellite imagery and field
plot data. The alternative we will research, called "k-nearest-neighbors," assigns
known characteristics of field sites (e.g., FIA plots or other plots/stands)
to the image elements that are their closest neighbors in spectral space.
Among other things, it has the potential to provide improved estimates
of the forest cover types and volumes for small areas of mixed ownership
where we may have good image data but little or no ground information and
FIA estimates are not very precise.
- Develop improved image product solutions for forest
management applications. Three specific field forestry problem areas where
satellite imagery has the potential for replacing current spatial data technologies
will be identified. For each problem area, optimal imagery combinations and
imagery expressions for visual interpretation will be developed and evaluated,
particularly in comparison to typical current day solutions. Delivery mechanisms
based on Internet technologies will be established that streamline the process
of getting the new products into the forester's hands.
The above objectives range from
strategic inventory to field level management tasks. To accomplish
them,
we have worked closely
with the USDA Forest Service North Central Research Station,
Minnesota DNR, the City of Eagan Forestry Division, Cass and
St. Louis County Land Departments, UPM – Kymmene Blandin
Paper, Potlatch Corp., and Pro-West, Inc. The Forest Service
and DNR are responsible for forest inventory in the state, while
Blandin and Potlatch are the two largest forest products companies
in Minnesota. Cass and St. Louis Counties manage 1.1 million
acres of county-owned forest lands. The City of Eagan Forestry
Division is a forerunner in applying geospatial technology to
urban resource management issues. Pro-West, Inc. supplies remote
sensing and GIS data and services to local governments and private
land managers. |
