Plant phenological records are crucial for predicting plant responses to global warming. However, many historical records are either short or replete with data gaps, which pose limitations and may lead to erroneous conclusions about the direction and magnitude of change. In addition to uninterrupted monitoring, missing observations may be substituted via modeling, experimentation, or gradient analysis. Here we have developed a space-for-time (SFT) substitution method that uses spatial phenology and temperature data to fill gaps in historical records. To do this, we combined historical data for several tree species from a single location with spatial data for the same species and used linear regression and Analysis of Covariance (ANCOVA) to build complementary spring phenology models and assess improvements achieved by the approach. SFT substitution allowed increasing the sample size and developing more robust phenology models for some of the species studied. Testing models with reduced historical data size revealed thresholds at which SFT improved historical trend estimation. We conclude that under certain circumstances both the robustness of models and accuracy of phenological trends can be enhanced although some limitations and assumptions still need to be resolved. There is considerable potential for exploring SFT analyses in phenology studies, especially those conducted in urban environments and those dealing with non-linearities in phenology modeling.
ASJC Scopus subject areas