U.S. national parks are often used as reference conditions for landscape change ecological studies. National park landscapes are thought to exhibit ecological conditions and landscape dynamics similar to those found prior to significant European human influence (Piekielek and Hansen 2012). This is despite little quantitative data describing national park landscapes even throughout the fairly recent past, like the last 50-100 years. What historic observational data do exist about national parks are often limited to notations of a single species, at a single point in time, at individual and disconnected field plots like those captured in present day museum specimens. Instead the ideal dataset would be spatially continuous, cover decades to centuries with multiple observations and be of fine spatial resolution relative to many common and landcover dominant species (i.e. not satellite imagery with 30-meter pixels like the original Landsat missions). Tree core data, gridded field plots from recent comprehensive national park inventory efforts, as well as contemporary geospatial landcover datasets all offer important insight, but deviate from the ideal landscape change dataset (Figure 1). Fortunately, there exists an underutilized geospatial data resource covering many national parks – archival (i.e. historic) aerial photography that when analyzed alongside contemporary aerial photography offers promise of exposing national park landscape dynamics for decades. Historic aerial photography often covers park lands in their entirety (i.e. is spatially continuous), has a spatial resolution of 1-meter or smaller, and goes back in time to sometimes the 1920s or 1930s often with multiple observations up to the present. Historic aerial photography has been used in a few studies to study forest structure and land cover change (Bozek et al. 2019) as well as geomorphic and land use change (Fuerer and Vinatier 2018). Historic aerial photography may be the best data source available to study landscape change in U.S. national parks.
Figure 1. Properties of landscape change datasets
Preventing the widespread use of historic aerial photographs in studies of landscape change is that they often exist in analog form as film or prints preserved in the inaccessible and out of view archives of government agencies. Furthermore, even once digitized, historic aerial photography in their native analog or digital form contain geometric displacements, distortions and other feature inaccuracies that prevent them from being incorporated directly into contemporary geospatial studies that require a uniform spatial scale that results from the process of orthorectification. Softwares to perform this transformation are becoming more common and national parks present a robust test of these tools due to their extreme mountain terrain and often continuously forested landscapes. The present project used Glacier and Crater Lake National Parks as tests of the latest orthorectification softwares from Agisoft and ESRI during a sabbatical research experience in spring 2022.
Glacier National Park contracted its own aerial photography in 1968, collecting over 1,500 black and white single frame images of park condition at that time. Crater Lake National Park was covered in a 1982 color-infrared aerial photography mission run by the U.S. Department of Agriculture. Once digitized, orthorectified, mosaicked and inspected relative to contemporary datasets, historic aerial photographs provided evidence of landscape changes of interest to scientists and park managers including conifer encroachment into lower elevations as a result of fire suppression (figure 2), changes in avalanche chute dynamics, the effects of large scale wildfire (figure 3), stream and river channel migration, evidence of low-density human development in the park periphery (figure 4), upslope movement of vegetation communities in response to climate change (figure 5), and lake status (i.e. frozen versus thawed) and extent, among other important ecological changes.
Figure 2. A natural color depiction of conifer encroachment into lower elevations as a result of fire suppression over the last 40 years in Crater Lake National Park. Initial conditions presented are 1982 and the image swipe presents landscape conditions in 2017.
Figure 3. Color infrared historic aerial photography show the effects of landscape scale wildfire in Crater Lake National Park. The initial conditions show a largely intact conifer forest in 1982 and the image swipe reveals the same forest almost completely gone by 2017.
Figure 4. This natural color historic aerial photography shows low-density human development near Crater Lake National Park from 1982 to 2017.
Figure 5. This transition from a 2019 natural color aerial photography to a 1968 black and white photograph shows the upslope migration of conifer forests in Glacier National Park over this time-period as a result of climate change.
Historic aerial photography presents an ideal data source with which to investigate landscape change and dynamics in U.S. national parks. That said, photographs from different flight missions and from different geographies present their own unique challenges to orthorectification. The deep canyons south and west of Crater Lake National Park exhibited photo shadowing that negatively affected the positional accuracy of the orthorectification result, whereas the steep rocky peaks of Glacier National Park did the same in that study domain and including mismatches at photograph seamlines.
Despite the challenges of working with historic aerial photography for landscape change studies, there exist few if any alternative quantitative datasets. Fortunately, the software tools to orthorectify historic aerial photographs continue to improve and become more user-friendly as well as perform better with respect to the horizonal positional accuracy and visual aesthetics of results.
References Cited.
1.Bożek, Piotr, Jaroslaw Janus, and Bartosz Mitka. “Analysis of Changes in Forest Structure Using Point Clouds from Historical Aerial Photographs.” Remote Sensing 11, no. 19 (September 27, 2019): 2259. https://doi.org/10.3390/rs11192259.
2.Feurer, D., and F. Vinatier. “Joining Multi-Epoch Archival Aerial Images in a Single SfM Block Allows 3-D Change Detection with Almost Exclusively Image Information.” ISPRS Journal of Photogrammetry and Remote Sensing 146 (December 2018): 495–506. https://doi.org/10.1016/j.isprsjprs.2018.10.016.
3.Piekielek, N.B. and A.J. Hansen. (2012). Extent of fragmentation of coarse-scale habitats in and around US National Parks. Biological Conservation 155:13-22.
Leave a Reply