Urban plant phenology monitoring: Expanding the functions of widespread surveillance cameras to nature rhythm understanding

Abstract

Phenology is an important ecological indicator for understanding the feedback of plants to climate changes, but observation of plant phenology is not a trivial task, particularly for the large-scale areas of interest. Urban plant phenology monitoring is such a typical case, since massive residents do not necessarily mean enough eligible phenology observers. To handle this traditional challenge, this study attempted those surveillance cameras (SCs) widespread almost in any city all over the world. The schematic plan is to install an automatic software module, which has the function of plant phenology monitoring, as a plug-in into any central unit that wire-controls RGB SCs. The kernel of the module is a general-purposed algorithm capable of deriving the starting and ending dates of the key phenological events of different plants that stand in the field of view of each telecontrolled SC. The kernels of the algorithm comprise deriving phenological indices from the digital number (DN) records by a SC from all of its RGB channels and, then, modeling of plant dynamics based on the proposed novel phase-limited multi-Gaussian model for curve-fitting of phenological phases. In the case of determining the key phenological dates regarding flowering and foliation in this study, tests suggested that the proposed scheme and phenological indices and the programmed software plug-in all worked well. Overall, the feasibility of using the widespread SCs for urban plant phenology monitoring was validated, and the scheme can be further extended to composing phenology observation networks at local or global scales. The solution is of implications for more understanding the interannual rhythms of terrestrial ecosystems as well as the inherent mechanisms of vegetation-climate interactions.