Integration of strategic inventory and monitoring programs for the forest lands, wood lands, range lands and agricultural lands of the United States
Czaplewski, R.L.
Proceedings Rocky Mountain Research Station, USDA Forest Service ( RMRS-P-12): 342-348
1999
Document Number: 337846
The United States Department of Agriculture uses the Forest Inventory and Analysis (FIA) program to monitor the nation's forests and woodlands, and the National Resources Inventory (NRI) program to monitor the nation's agricultural and range lands. Although their measurement methods and sampling frames are very different, both programs are developing annual systems to better detect trends in land use and ecosystem health, evaluate effectiveness of public policies, guide sustainable development, and forecast alternative future conditions. Other federal programs use Landsat satellite data to map the nation's land cover, land use, and habitat diversity. I offer two proposals that could increase consistency and reduce cost among these federal monitoring programs. First, a national consortium would acquire Landsat data every two to five years for the entire USA, and rapidly detect abrupt changes in spectral reflectance associated with land clearing and major changes in land cover. This would provide spatial data to update existing maps of land cover and land use, and improve statistical monitoring. However, Landsat resolution is not sufficient to identify detailed categories of land use and forest cover; rather, higher-resolution sensors are necessary to significantly reduce the required amount of field data. Therefore, the second proposal is acquisition and interpretation of large-scale high-resolution aerial photography for hundreds of thousands of FIA and NRI sampling units. This ambitious enterprise might be feasible with collaboration between these two programs. They would integrate their separate photo-interpretation operations, and collocate 65-ha NRI sampling units with 1-ha FIA field plots. Each year, a 20% sub-sample of the collocated sampling units is photographed from low-elevation aircraft to detect rapid and obvious changes, while field data are gathered for a 10% sub-sample of field plots to detect slower and more subtle changes. Field data are paired with remotely sensed data, and empirical models statistically correct for measurement and classification errors in remotely sensed data. Implementation is a major enterprise that requires unprecedented partnerships among federal programs. I do not believe remote sensing will be an effective technological solution without such an enterprise. Most legislators and executive leaders are not aware of the magnitude of the logistical and institutional challenges. However, extensive coordination of remote sensing among existing federal programs could produce an efficient system to evaluate sustainability of the nation's forests and agricultural lands, and assess the health of the nation's ecosystems.