Ecological Archives E092-043-A3

Kirk M. Stueve, Rachel E. Isaacs, Lucy E. Tyrrell, and Roseann V. Densmore. 2011. Spatial variability of biotic and abiotic tree establishment constraints across a treeline ecotone in the Alaska Range. Ecology 92:496–506.

Appendix C. Expanded methods for digital data acquisition and processing.

Aerial Photography

The United States Geological Survey (USGS) supplied unprojected panchromatic aerial photography from July 1953 at a spatial resolution of ~1.5 m. There were unfortunately little metadata available for the photography to supplement orthorectification procedures (e.g., flight and camera specifications).

Aerial Photography

Acquisition Date

25 July 1953

Spatial Resolution (m)

~1.5

Flight Altitude (m)

6096

Focal Length (mm)

153.69

Resolution (dpi)

1200

Image Size (cm)

22.9

Image Type

Panchromatic

Agency

US Airforce

 

IKONOS Satellite Imagery

The National Park Service (NPS) supplied orthorectified pan-sharpened IKONOS satellite imagery from July 2005. The imagery has a spatial resolution of 1 m and horizontal accuracy of +/- 15 m, which are similar to the specifications of most USGS DOQs widely used by ecologists as base imagery in a GIS. IKONOS panchromatic imagery has a spatial resolution of 0.82 m, whereas the multispectral imagery has a spatial resolution of 4 m. The process of combining these two images with pan-sharpening procedures results in imagery with a spatial resolution of 1 m.

IKONOS

Acquisition Date

15 July 2005

Spatial Resolution (m)

1

Orbital Altitude (km)

681

Swath Width (km)

11.3

Image Type

Pan-sharpened

Agency

GeoEye

 

Digital Elevation Model

The USGS supplied a 60-m digital elevation model (DEM). We preferred a DEM with a finer spatial resolution and improved accuracy, but potential options for DEMs are limited in the remote locations of Interior Alaska. Nonetheless, the 60-m DEM generally has acceptable vertical and horizontal accuracies of at least +/- 30 m and +/- 30 m respectively. We applied standard processing procedures to the 60-m DEM with the ArcInfo 9.2 software package from ESRI (200×) in order to fill several sinks and smooth out any remaining systematic errors (e.g., Stueve et al. 2009). Then, we used the flow direction and flow accumulation functions in ArcInfo 9.2 to delineate stream networks.

Image Processing Procedures

We geometrically tied the stream networks generated from the DEM with ArcInfo 9.2 to the 2005 IKONOS imagery to ensure precise alignment in a GIS. We used a 2nd order polynomial transformation (superior in mountainous terrain) and 35 points evenly distributed across the imagery (including edges) to obtain a root mean squared error (RMSE) below 0.5 pixels (< 30 m) (Jensen 2005). Then we subjected the 1953 aerial photography to orthorectification procedures and used the 2005 IKONOS image as a reference to collect ground control points. We were only able to employ a topographic correction by using the DEM to model topography because necessary metadata for flight and camera specifications were not available. To compensate for this lack of metadata and associated photography distortions resulting from varying angles between the camera in the airplane and disparate ground locations, we performed a series of orthorectification procedures at restricted spatial extents (i.e. 400 × 400 m) across the study site. This enabled us to achieve RMSEs near or below one pixel in every case (~1.5 m or less). In each series, we selected 25 widely distributed control points with several placed along the periphery of the photography, and used orthorectification procedures similar to those discussed by Stueve et al. (2009). We applied the same orthorectification procedures to the entire study area and could not lower the RMSE below 18 pixels. In this case, visual observations indicated some of the high elevations (i.e., > 1500 m) displayed horizontal errors exceeding 50 m. Conversely, it was difficult to discern any discrepancies between the 1953 photography and 2005 imagery when comparing the series of orthorectifications (Figure 3 from the main text). Thus, we are confident in the accuracy of our data at both the upper and lower altitudinal extremes.

LITERATURE CITED

Jensen, J. R. 2005. Introductory Digital Image Processing: A Remote Sensing Perspective. Third edition. Prentice Hall, Upper Saddle River, New Jersey, USA.

Stueve, K. M., D. L. Cerney, R. M. Rochefort, and L. L. Kurth. 2009. Post-fire tree establishment patterns at the alpine treeline ecotone: Mount Rainier National Park, Washington, USA. Journal of Vegetation Science 20:107–120.


[Back to E092-043]