The Phenological Parameters Estimation Tool (PPET) is a set of algorithms implemented in MATLAB that estimates key vegetative phenological parameters. For a given year, the PPET software package takes in temporally processed vegetation index data (3D spatio-temporal arrays) generated by the time series product tool (TSPT) and outputs spatial grids (2D arrays) of vegetation phenological parameters. As a precursor to PPET, the TSPT uses quality information for each pixel of each date to remove bad or suspect data, and then interpolates and digitally fills data voids in the time series to produce a continuous, smoothed vegetation index product. During processing, the TSPT displays NDVI (Normalized Difference Vegetation Index) time series plots and images from the temporally processed pixels. Both the TSPT and PPET currently use moderate resolution imaging spectrora-diometer (MODIS) satellite multispectral data as a default, but each software package is modifiable and could be used with any high-temporal-rate remote sensing data collection system that is capable of producing vegetation indices.

Raw MODIS data from the Aqua and Terra satellites is processed using the TSPT to generate a filtered time series data product. The PPET then uses the TSPT output to generate phenological parameters for desired locations. PPET output data tiles are mosaicked into a Conterminous United States (CONUS) data layer using ERDAS IMAGINE, or equivalent software package. Mosaics of the vegetation phenology data products are then reprojected to the desired map projection using ERDAS IMAGINE.

The TSPT software can temporally process a variety of MODIS multispectral data products on a per-pixel basis. Generally, the TSPT runs on one user-specified kind of MODIS data product to generate a given time series data product. The TSPT can process output from the MODIS Re-projection Tool (MRT) as input, or can directly convert MODIS Hierarchical Data Format (HDF) sinusoidal gridded data to BSQ input files.

Unlike other known vegetation phenological parameter estimation software, the PPET produces not only common phenological parameters, but also real-time and custom parameters without a priori assumptions about the shape of the phenological cycle. Common phenological parameters, like those produced in PPET, are associated with the annual vegetation growth cycle. They quantitatively describe vegetative states related to annual cyclical growing seasons, such as green-up, maturity, senescence, and dormancy by analyzing the temporal shape of given vegetation index time series. The real-time phenological and custom parameters are formed from a cumulative sum (integral) produced at a fixed temporal interval. In addition, cumulative vegetation index and time-specific/pest-specific phenological parameters can be designed to optimize the detection of vegetation damage from specific pests and diseases. These problem-specific, phenological parameters have the potential to be integrated into near real-time, predictive surveillance systems (i.e. early warning systems) and, with improved vegetative state information, could assist decision makers in making intelligent vegetation and associated land resource management choices.

This work was done by Rodney D. McKellip of Stennis Space Center; Kenton W. Ross, Joseph P. Spruce, James C. Smoot, and Robert E. Ryan of Science Systems and Applications, Inc.; Gerald E. Gasser of Lockheed Martin; and Donald L. Prados and Ronald D. Vaughan of Computer Sciences Corp.

Inquiries concerning rights for the commercial use of this invention should be addressed to

the Intellectual Property Manager at Stennis Space Center (228) 688-1929.

SSC-00321