Phenology is an important way to observe impacts of climate change on ecosystems. Long-term, landscape-scale phenology monitoring has gained substantial interest in the last 25 years, largely due to observed changes in the timing of growing season events. Landscape-level studies of phenology present unique challenges. Although direct measurements of phenology are now made across the country and coordinated by the USA National Phenological Network, it is difficult to extrapolate local observations across large regions. Indirect inferences of landscape-level phenological events can be made utilizing optical remote sensing observations based on time-series data (e.g., NDVI). A geospatial database for the conterminous United States of nine phenological measures depicted at two different resolutions (250 m² and 1,000 m²) for over two decades is accessible online (http://phenology.cr.usgs.gov/) and includes dates for the beginning and end of each growing season. The AVHRR and MODIS sensors provide input data for this phenology database. The database has been used for research into the phenological patterns of aspen woodlands, to help model the spread of invasive species (e.g., cheatgrass), and to model and monitor seasonal fire danger across the United States. In the aspen study (image below), the relationship between the start of the season and the spring snow water equivalent was found to be strong and significant (R² = 0.84, p0.001).
Ten-year time series of annual start of growing season for aspen woodlands in southwest Colorado (38º N, 108º W).