MODIS Land Cover GroupDepartment of Geography and Environment at Boston University
Land Cover and Land Cover Dynamics - Classifying global biomes and monitoring their dynamics ...
The MODIS Land Cover Dynamics Product (MOD12Q2)
The MODIS land cover product is designed to support scientific investigations that require information regarding the state and seasonal-to-decadal scale dynamics in global land cover. The product consists of two suites of science data sets (SDS’s). MODIS land cover type (MOD12Q1), includes five main layers in which land cover is mapped using different classification systems. MODIS land cover dynamics (MOD12Q2) includes seven layers, and has been developed to support studies of seasonal and interannual variation (phenology) in land surface and ecosystem properties. Both products are global. In collections 1, 3 and 4 MOD12 was produced at a spatial resolution of 1-km. In collection 5, the spatial resolution has been increased to 500-m.
The MOD12Q2 product provides global maps of seven phenological metrics for all ecosystems exhibiting identifiable annual phenologies. These metrics include the date of year for: (1) the onset of greenness increase (greenup), (2) the onset of greenness maximum (maturity), (3) the onset of greenness decrease (senescence), and (4) the onset of greenness minimum (dormancy). The three remaining metrics are the growing season minimum, maximum, and summation of modeled daily vegetation index values from MODIS.
Algorithm Description
The MODIS land cover dynamics algorithm provides a remote sensing-based methodology for identifying transition dates that define the key phenological phases of vegetation at annual time scales (Zhang et al., 2003). To estimate these transition dates, EVI values computed from NBAR data are used. Because the presence of snow can significantly affect EVI values, the EVI data are first screened to remove (or at least minimize) snow contaminated pixels. Specifically, in cases where snow or ice is detected, the EVI value is replaced with the most recent snow-free value.
The figure at right provides a schematic representation for how vegetation phenology is modeled by the MOD12Q2 algorithm using series of piecewise logistic functions of time. Specifically, temporal variation in satellite derived EVI data for a single growth or senescence cycle is modeled using a sigmoidal function, which is fit to time series of EVI data using least squares. To identify phenological transition dates, the rate of change in the curvature of the fitted logistic models is used. Transition dates correspond to times at which the rate of change in curvature in the EVI data exhibits local minima or maximums.
During the growth period, when vegetation transitions from a dormant state to maximum leaf area, three extreme points in a EVI curve can be idenitified. These points correspond to the onset of leaf growth, the onset of maximum leaf area, and the inflection point between these two events. Transition dates indicating the onset of senescence and dormancy can be estimated in a similar fashion. Note that before the NBAR EVI data are fit using logistic functions, it is first necessary to identify periods of sustained EVI increase and decrease (i.e., growth and senescence). To do this, a moving window using five 16-day periods is applied to the annual time series. Transitions from increasing to decreasing EVI trends are identified by a change from positive to negative slope within any given window, and vice versa. In this way, distinct growth cycles can be identified within a given annual time series.