The use of PlanetScope CubeSats for forestry : a literature review and a case study

Constellations of Earth observation satellites have become a critical component of forest management due to their ability to acquire data over large spatial areas at regular intervals. First launched in 2014, PlanetScope is the largest such constellation currently in operation. By using 100–200 CubeSats—small satellites made of relatively inexpensive standardized components— PlanetScope is able to provide high-resolution daily imagery of the entire surface of the Earth, the first satellite system to do so. These qualities make PlanetScope a promising technology for near-real time forest monitoring or for monitoring changes in forests at fine spatial scales. However, because the CubeSat sensors have lower levels of cross-calibration when compared with other satellite constellations, many users have reported challenges with inconsistent quality in PlanetScope data. To understand the relative strengths and limitations of PlanetScope data for forest monitoring, I completed a systematic literature review of more than 150 relevant peer reviewed publications. The findings of this review indicate that while the high spatial and temporal resolutions of PlanetScope are beneficial in many contexts, uptake of data is potentially limited by issues with its quality and consistency. Many methods of normalizing PlanetScope data have been developed; however, these methods remain underutilized, likely due to their complexity and lack of availability in public software packages. To address this barrier, I subsequently completed a case study to test whether several simple statistical transformations could be used to normalize time series of PlanetScope data for detecting fine-scale forest disturbances. I found that two transformations—the Z-score and robust Z-score—were effective for denoising the data and suppressing the phenological signal for several stands in British Columbia, Canada, thus, areas that had undergone thinning were more easily differentiated from areas that had not undergone thinning. Overall, these findings suggest that with appropriate and accessible data processing methods, PlanetScope has strong potential to support reliable, fine-scale forest monitoring in both research and operational management contexts.