New ecological change indicators using breakpoints in vegetation trends applied to a dryland catchment in Morocco’s High Atlas

Change detection in vegetation index time series can identify abrupt responses to disturbances and assess ecosystem stability. Breakpoints in vegetation greening trends are being used for this purpose, with more breakpoints indicating lower resistance. However, breakpoints can be positive or negative, reflecting improving or degrading trends. Therefore, areas with similar breakpoint counts may differ in stability depending on the balance of positive and negative changes. This study investigates how incorporating breakpoint sign improves the assessment of vegetation dynamics and develops an improved typology based on the sign and significance of trend slopes before and after breakpoints. We applied the new breakpoint typology to 35 years of Landsat NDVI data from a pastoral catchment in Morocco’s High Atlas. We derived the total, positive and negative number of breakpoints in NDVI trends accumulated during the study period and the type of breakpoint in response to the most severe drought within that period. Regions with smaller NDVI changes over time exhibited a higher number of breakpoints with a similar share of positive and negative, compared to areas with stronger greening/ browning, which a higher share of positive/negative breakpoint. During the drought, positive breakpoints (positive reversals) were most common, followed by negative breakpoints (interrupted decreases). Areas with positive reversals experienced fewer total breakpoints over the study period and had a greater share of positive breakpoints than areas with interrupted decreases. These findings highlight the importance of analysing the balance of positive and negative breakpoints alongside their total count for understanding ecological change.