The relationship between climate and satellite-derived vegetation phenology

Phenology, the timing of recurring biological life cycle events, is usually influenced by photoperiod, precipitation, soil and air temperature, and other life-controlling factors. Satellite remote sensing has become an essential tool for measuring and monitoring large-area phenological variability and can reveal broad-scale phenological trends that would be difficult to detect from ground observations. A 20-year (1989-2008) historical phenological database for the conterminous United States calculated from 1-km Advanced Very High Resolution Radiometer (AVHRR) satellite time series data provides a powerful tool for documenting and understanding phenological events over time and detecting the impacts of climate change on ecosystems. In this study, we assess important features in the relationship between climate conditions and vegetation phenology. The data include several phenological variables from the USGS database [start of season time (SOST), end of season time (EOST), and time integrated normalized difference vegetation index (TIN)] and two basic climate variables [PRISM (Parameter-elevation Regressions on Independent Slopes Model) monthly precipitation and temperature]. The investigation focuses on several study areas representing different ecological regions in the United States. Results for one study area, the Tallgrass Prairie National Preserve site in Chase County, Kansas, revealed relationships between these climate and phenological variables. For this site, there were clear relationships among monthly minimum temperatures (Tmin) during certain months and seasonal phenological events (i.e., SOST and EOST). The highest monthly Tmin (3 °C) at the start of season was found in March 2004, which corresponded to the earliest SOST (March 30). Likewise, the lowest March Tmin (-3 °C) was in 2002, which had the latest SOST (April 18). Similarly, the highest monthly Tmin (4 °C) near the end of season was found in November 2001, which had the latest EOST (November 16), while the lowest November Tmin (-3 °C) was in 2000, the year with the earliest EOST (October 31). Our results also exhibited a relationship between annual precipitation (AP) and TIN. The years (1997-1999, 2001-2004) with above average AP (> 787 mm) had much higher TIN (43-65) than the years (2000 and 2002) with below average AP (TIN from 40 to 44). Results from this study will provide useful information on how climate changes affect phenology, which would be used for future ecosystem management.