Climate change and its effects on vegetation phenology across ecoregions of Ethiopia

Abstract

Vegetation phenology is an integrative environmental indicator of climate change and long-term observations of the changes in plant phenology using remote sensing technologies help us to understand climate change trends over space and time. However, such trends and their implications for ecosystem health have been poorly explored in Ethiopia. In this paper we examine the temporal changes in the phenology of vegetation in relation to climatic drivers across Ecoregions in Ethiopia using satellite images. To do this, the MODIS 8-day NDVI product, MODIS surface temperature and emissivity, and pentad based rainfall data from Climate Hazards Group InfraRed Precipitation with Station (CHIRPS) for the 14 year period 2002–2015 were used. The aggregated time series NDVI, temperature and rainfall data were generated for each Ecoregion in the Google Earth Engine (GEE) Environment followed by Fourier smoothing to overcome various noises. The phenology of each Ecoregion was constructed using the intra-annual NDVI variability. Major events in the vegetation cycle or phenophases were identified using sigmoid vegetation growth functions and inflection point detection techniques. The relationships between rainfall and NDVI and between temperature and NDVI were investigated and multiple regression models were developed as regressors to NDVI. Our results indicated that, over the 14 years, the start of the growing period became earlier, and the growing period elongated for most of the Ecoregions in Ethiopia, except for the Somali Acacia-Commiphora bushlands and thickets (SACBT) Ecoregion, in which the start of the growing period was delayed. The widening of the growing period indicates an increase in ecosystem productivity, increment of evapo-transpiration, and disturbance to the water and energy balance of the region. Rainfall and NDVI were positively and strongly correlated but with one-month lag time, whereas temperature and NDVI were negatively correlated for all the Ecoregions. This appears to be due to the variation in timing of the high rainfall season and the decreasing trends of temperature with respect to the variation in the angle of the sun and the subsequent movements of the ITCZ in the region. The overall study indicates that climate variability is affecting the phenology of vegetation across all Ecoregions in Ethiopia. Shifts in crop growing seasons should be considered to efficiently utilize the summer rain for crop production. Keywords: Phenology, Phenophase, NDVI, Rainfall, Temperature, Ecoregions