Classifying Timeseries Trajectories
We're utilizing the full archive of Landsat data, available through Google Earth Engine, to characterize changes in vegetation in northern China over the past 40 years. We then utilize this information to create a new classification based on similarities in timeseries signatures, such as the timing and magnitude of changes. These unique archetypes can be linked to specific policy and management interventions.
Grad student Mia Murray is developing this work to map land cover histories in the Xilingol grasslands of Inner Mongolia
Establishing new baseline data on historic land cover in northern China
Read more about this project in my tweet thread here!
Most of the information we have about land cover change over time is limited temporally by the availability of satellite data, so most work only extends to the early 1980s. But what if we could go back further??
Grad student Brooke Iacone has georeferenced a set of declassified surveillance imagery from northern China in the late 1960s to give us our first understanding of the scale of grassland degradation during the Cultural Revolution. She is using this information to generate a land cover classification for 1970,