The objectives of this study were to evaluate the utility of growing season parameters (beginning, end, and length) derived from satellite sensors, by comparing them with those determined through surface phenological analyses at local scales, and to verify reliability of extrapolating growing season parameters using threshold normalized difference vegetation index (NDVI) values at regional scales.

First, satellite sensor-derived growing season parameters were calculated using NDVI time series and statistical procedures at three sample stations in the temperate region of eastern China, from 1982 to 1993. Then, a comparative evaluation of these growing season parameters was made using actual phenological growing season parameters of local plant communities at each station over the same period. Significant correlations were found between the two types of growing season parameters in the middle and southern parts of the research region.

Moreover, beginning and end dates of the surface phenological growing season were applied as time thresholds to determine corresponding 10-day peak greenness values from NDVI curves at pixels overlying these stations. Using the threshold NDVI values from individual years, we carried out a spatial extrapolation to examine whether the thresholds were appropriate for determining growing season parameters at adjacent sites within the research region. Initial results show that corresponding threshold NDVI values at northern sample stations have stronger extrapolation capability than at southern stations, and extrapolating beginning dates of the growing season should be more reliable than end dates.

In an on-going phase of the work, statistical relationships will be established to monitor and estimate the surface phenological growing season based on NDVI data at local and regional scales.