Extraction of Phenological Metrics for Mixed Deciduous Forest of Doon Valley, India using MODIS NDVI Time Series Data

Evaluating spatiotemporal changes in forest phenology can help understand its response to the climatic drivers and its linkage with ecological processes. In the current study, a total of 230 datasets ( 20092018) of 16-day MODerate Resolution Imaging Spectroradiometer (MODIS) MOD13Q1 time series Normalized Difference Vegetation Index (NDVI) product were used for the extraction of the phenological metrics for the mixed deciduous forest in Doon valley, India. Primary phenological metrics include the start of the growing season (SoS), end of the growing season (EoS), the peak of the growing season (PoS), length of the growing season (LoS), seasonal amplitude, large and small integral. The first step includes the conversion of the datasets into the binary format. An adaptive SavitzkyGolay filtering function was used in the second step to fit the curves and smoothen the time series dataset. A median filter was used for the removal of the spikes and outliers in the dataset. The seasonality parameter value was set as 0.5, and the threshold distance was set to 20% for the beginning and end of the season. All the processes were executed using TIMESAT software. Results showed that SoS varied from 176 to 209 day of the year (DOY) (190 ±13), and EoS varied from 458 to 489 DOY (475±9). LoS ranges from 266 to 303 DOY (285±13). PoS reaches in September and October (273±15) with a mean NDVI value of 0.83. The lower value of small integral (5.68 ± 0.65) and the higher value of large integral (12.58 ± 0.63) with a mean amplitude of 0.48 signify the high productivity of the forest. There is no distinct pattern observed between the rainfall and the phenological metrics, which can be attributed to the high variability of the rainfall in the valley.