DIFFERENCES IN ONSET OF GREENNESS: A MULTITEMPORAL ANALYSIS OF GRASS AND WHEAT IN KANSAS*

The time when vegetation emerges (begins active photosynthesis) each year can be monitored using remotely sensed data obtained from Earth observation satellites. This measurement, called the onset of greenness, can be calculated using time series data sets of the Normalized Difference Vegetation Index (NDVI), a ratio of red and near infra-red (NIR) light that is strongly correlated to plant biomass. The date at which the onset of greenness occurs varies depending on the region, type of vegetation present, weather conditions, land management practices, and other factors. Time series NDVI data are proving useful for a wide variety of applications including vegetation mapping, crop monitoring, and yield modeling. The onset of greenness metric values were extracted from bi-weekly AVHRR satellite data over the state of Kansas from 1989 through 2000. Four different locations in Kansas were sampled: 1) winter wheat from Sumner County, 2) winter wheat from Thomas County, 3) tall grass rangeland from Chase County, and 4) short grass rangeland from Logan County. The onset date was recorded from 20 AVHRR pixels from each location for each year of data. Samples from the same location were analyzed to detect differences in onset date between years and samples from different locations were compared to identify differences in onset date within the same year. Results show a significant difference in the onset of greenness date between cover types and a significant difference in onset dates between years for the same cover type at the same location. Regression analysis shows that the grasslands and wheat of northwestern Kansas have a negative slope (earlier onset), while the lines of best fit for grasslands and wheat in south central Kansas did not indicate a substantial change in onset date.