Near real-time disturbance detection in terrestrial ecosystems using satellite image time series : drought detection in Somalia

Near real-time monitoring of ecosystem disturbances is critical for addressing impacts on carbon dynamics, biodiversity, and socio-ecological processes. Satellite remote sensing enables cost-effective and accurate monitoring at frequent time steps over large areas. Yet, generic methods to detect disturbances within newly captured satellite images are lacking. We propose a generic time series based disturbance detection approach by modelling stable historical behaviour to enable detection of abnormal changes within newly acquired data. Time series of vegetation greenness provide a measure for terrestrial vegetation productivity over the last decades covering the whole world and contain essential information related land cover dynamics and disturbances. Here, we assess and demonstrate the method by (1) simulating time series of vegetation greenness data from satellite data with different amount of noise, seasonality and disturbances representing a wide range of terrestrial ecosystems, (2) applying it to real satellite greenness image time series between February 2000 and July 2011 covering Somalia to detect drought related vegetation disturbances. First, simulation results illustrate that disturbances are successfully detected in near real-time while being robust for seasonality and noise. Second, major drought related disturbance corresponding with most drought stressed regions in Somalia are detected from mid 2010 onwards and confirm proof-of-concept of the method. The method can be integrated within current operational early warning systems and has the potential to detect a wide variety of disturbances (e.g. deforestation, flood damage, etc.). It can analyse in-situ or satellite data time series of biophysical indicators from local to global scale since it is fast, does not depend on thresholds or definitions and does not require time series gap filling. -- early warning ; real-time monitoring ; global change ; disturbance ; time series ; remote sensing ; vegetation and climate dynamics