Pixel-wise abrupt change point detection for raster time series using Pettitt's non-parametric test
TrendShift detects abrupt temporal shifts in time-ordered raster stacks. It estimates the most likely change point per pixel using Pettitt's non-parametric change point test and exports change step, change time, magnitude, direction, p-value, significance mask, and test statistic rasters. Designed for NDVI shifts, rangeland stress, forest disturbance, land-cover transitions represented as continuous indices, precipitation/temperature anomalies, water extent metrics, and other environmental raster time-series workflows.
Plugin Tags