An investigation of observed phase differences between clearcut and standing forests in interferograms from the Pacific Northwest

Since the early 1990's, interferometric synthetic aperture radar (InSAR) has proven to be a useful technique for observing sources of crustal deformation. As data has become more widely available, much work has been done to understand tectonic deformation signals observed in interferograms. In order to more fully understand these tectonic signals, it is necessary to also study signals associated with non-tectonic sources. This study addresses a non-tectonic source of phase changes associated with extensive clearcutting of forests in the Pacific Northwest. Clearcutting has been utilized as a logging technique throughout the Pacific Northwest since the earliest days of the timber industry in the region. Areas that have undergone clearcutting during the past 20 years are associated with a coherent signal with a different line-of-sight phase change than the surrounding standing forest. In some instances, phase changes show discrete jumps correlated with multiple episodes of clearcutting in a region. There are a number of possible causes for the observed phase difference between the clearcut and standing forests. One explanation could be an effective digital elevation model (DEM) error, since we use a DEM in our interferometric processing from the Shuttle Radar Topography Mission (SRTM) that was generated using C-band (~6cm) wavelengths, while the ALOS data used in this study was acquired at L-band (~24cm). This effect would be expected to correlate with the length of the spatial baseline between the two satellite acquisitions. Tree growth that occurred between the SRTM and ALOS acquisitions may also explain the observed phase differences, as long as there was no significant growth between the two ALOS acquisition times used in each interferogram (which would otherwise cause decorrelation). Regions where several periods of clearcutting have occurred show multiple phase differences, with the most pronounced differences associated with the more recent episodes of clearcutting relative to the SAR acquisitions. This effect would also correlate with baseline, but also, potentially, with time. Soil moisture may also account for the observed difference in returned phase. Changes in soil moisture between areas where forests remain standing and where clearcutting has occurred may effect the dielectric constant of the soils, thus leading to a difference in phase between adjacent clearcut and standing forests. We would expect this relationship to be random in space and time. The temporal and spatial relationship between the phase differences of adjacent clearcut and standing forests will indicate which of the aforementioned explanations have caused the observed signal. Specifically, the baseline dependency of the adjacent phase differences must be determined. Comparisons of these clearcut signals with records of logging in the area will allow for identification of any trends in phase differences that correlate with distinct stages of clearcutting. Ultimately we hope to provide an explanation for a prominent non-tectonic signal found in interferograms of the Pacific Northwest, which may improve our interpretation of tectonic signals and may also have applications to monitoring tree growth in clearcut regions.