The strong relationships between vegetation phenology and global climate change have been
found in recent years, especially with increasing popularity and availability of satellite data.
Accurate estimates of canopy phenology are critical to quantify carbon and water exchange between
forests and the atmosphere and its response to climate change. The objective of this study is to detect
the spatial distribution of vegetation phenology with remote sensing and to quantitatively examine
the linkage between forest phenology and forest type in contiguous United States. In particular, we
focus on phenology variation between different forest types. To achieve this goal, we utilize LAI
measurements from Moderate Resolution Imaging Spectroradiometer (MODIS) from 2000 to 2007
to identify phenological transition dates. The transition dates are then related to MODIS land cover
type product to assess land cover type dependent phonological variation during 8 years. The results
show that both evergreen forests and deciduous forests have an annual cycle of vegetation phenology.
Greenup onset days vary diversely among different forest types. The phenology variation range of
deciduous needle leaf forests is larger than that of deciduous broadleaf forests. Compared to greenup
days, dormancy days have a little difference between different forest types. Grow length of different
land cover varies obviously during 8 years.
The wildfires which occurred in April 2007 in southern Georgia lasted for almost two months. Approximately 386,722 acres were burned. In this paper, we explored the strategy to use MODIS products to study fire impacts on vegetation. Vegetation variations caused by the fires are studied using these MODIS products: 8-day composite fire products, 16-day composite vegetation indices, and land cover types. Several tiles of MODIS products from dates immediately before and after the fire were employed to monitor vegetation changes. Unburned control plots were selected to establish new variables QNDVI and QEVI which are used to evaluate the vegetation recovery status. The results show that vegetation indices: NDVI and EVI decreased dramatically after the fires. The variations of QNDVI and QEVI indicate that vegetation status still has a disparity between fire spots and surrounding undisturbed area, even though commonly used vegetation indices of fire spots have attained pre-fire levels. The appropriate distance outside the fire spots for selecting control plots is related to the size of the burned area. The larger the burned area is, the bigger distance we should choose for control plots. EVI and QEVI are better indicators to show vegetation changes due to fires. The method presented in this paper can be employed to monitor vegetation change to fires as well as to indicate different recovery rate. It also can be useful to identify the fire severity and assess the ecological consequence of fires.