Dr Zhu Xiaolin, Member of the Research Institute for Land and Space (RILS), and his former PhD student Dr Tian Jiaqi published a comprehensive review on remote sensing of land surface phenology in Global Change Biology, a top journal in the field of ecology (IF: 13.211), in collaboration with scholars from inland China, Australia, Switzerland, the USA, and Singapore. They presented the key issues in the remote sensing of land surface phenology, that is, the monitoring of seasonal variation in vegetated land surface use by means of satellite and observation networks. These issues include the impact of spatial resolution, the viewing solar zenith angle, the temporal resolution, and the noise in the time series in vegetation phenology detection.
For the impact of spatial resolution, the team discovered that coarse-resolution satellite images (images with large spatial resolution) overestimate the rural–urban difference in phenological metrics. This overestimation can be attributed to the amplification of the rural–urban difference in spring vegetation phenology associated with spatial resolution. Since the majority of urban pixels in coarser images have higher diversity, the spring phenological dates (the start of the spring season) predicted from the images are earlier than the actual dates.
For the viewing solar zenith angle, existing studies have found that the seasonal changes in solar zenith angle (SZA), the angle between the sun’s ray and the vertical direction, can alter the temporal trajectory of the vegetation index (VI) time series, thereby reducing the precision of vegetation phenology extraction in comparison to that of fixed SZA.
For the temporal resolution, a recent study used the simulated enhanced vegetation index (EVI) with temporal resolutions of 1 day to 52 days to detect spring phenology in North America, and found that temporal resolutions nonlinearly affected the accuracy of land surface phenology (LSP).
For the noise in the time series, the team investigated the impact of clouds on the smoothing of satellite data for global-scale land surface phenology and recommended optimal smoothing parameters for future studies in different regions.
