Jump to Main Content
Assessing the performance of smoothing functions to estimate land surface phenology on temperate grassland
- Lara, Bruno, Gandini, Marcelo
- International journal of remote sensing 2016 v.37 no.8 pp. 1801-1813
- aerosols, computer software, cropland, ecosystems, grasslands, growing season, land use, models, normalized difference vegetation index, phenology, remote sensing, time series analysis
- NDVI (Normalized Difference Vegetation Index) time-series have been used for permitting a land surface phenology retrieval but these time series are affected by clouds and aerosols, which add noise to the signal sensor. In this sense, several smoothing functions are used to remove noise introduced by undetected clouds and poor atmospheric conditions, but a comparison between methods is still necessary due to disagreements about its performance in the literature. The application of a smoothing function is a necessarily previous step to describe land surface phenology in different ecosystems. The aims of this research were to evaluate the consistency of different smoothing functions from TIMESAT software and their impacts on phenological attributes of temperate grassland – a complex mosaic of land uses with natural vegetated and agricultural regions using NDVI-MODIS time series. An adaptive Savitzky–Golay (SG) filter, Asymmetric Gaussian (AG) and Double Logistic (DL) functions to fitting NDVI data were used and their performances were assessed using the measures root mean square error (RMSE), Akaike Information Criterion (AIC), Bayesian Information Criterion (BIC) and bias. Besides, differences on the estimation of the start of the growing season (SOS) and the length of the growing season (LOS) were obtained. High and low RMSE over croplands and grassland were observed for the three smoothing functions; in the rest of the region, the SG filter showed more reliable results. Patterns of difference on the estimation of SOS and LOS between SG filter and the other two models were randomly distributed, where differences of 20–50 days were found. This study demonstrated that methods from TIMESAT software are robust and spatially consistent but must be carefully used.