Authors: |
Raymond L. Ph.D..Czaplewski, Paul L. Patterson |
Year: |
2001 |
Type: |
Paper |
Station: |
Rocky Mountain Research Station |
Source: |
In: Reams, Gregory A.; McRoberts, Ronald E.; Van Deusen, Paul C., eds. 2001. Proceedings of the second annual Forest Inventory and Analysis symposium; 2000 October 17-18; Salt Lake City, UT. Gen. Tech. Rep. SRS-47. Asheville, NC: U.S. Department of Agriculture, Forest Service, Southern Research Station. pp. 32-42 |
Abstract
We specify accuracy standards for remotely sensed classifications used by FIA to stratify landscapes into two categories: forest and nonforest. Accuracy must be highest when forest area approaches 100 percent of the landscape. If forest area is rare in a landscape, then accuracy in the nonforest stratum must be very high, even at the expense of accuracy in the forest stratum. Accuracy in both strata must be at least 90 percent to achieve appreciable gains in efficiency. We recommend that new remotely sensed data be used to re-stratify landscapes whenever the area in forestland decreases by five percent or more since the previous stratification. Efficiency can increase up to 15 percent with formation of an "indeterminate" stratum, which contains elements that are most likely to be misclassified.
Citation
Czaplewski, Raymond L.; Patterson, Paul L. 2001. Accuracy of Remotely Sensed Classifications For Stratification of Forest and Nonforest Lands. In: Reams, Gregory A.; McRoberts, Ronald E.; Van Deusen, Paul C., eds. 2001. Proceedings of the second annual Forest Inventory and Analysis symposium; 2000 October 17-18; Salt Lake City, UT. Gen. Tech. Rep. SRS-47. Asheville, NC: U.S. Department of Agriculture, Forest Service, Southern Research Station. pp. 32-42