Food insecurity continues to affect more than two-thirds of the population in sub-Saharan Africa (SSA), particularly those depending on rain-fed agriculture. Striga, a parasitic weed, has caused yield losses of cereal crops, immensely affecting smallholder farmers in SSA. Although earlier studies have established that Striga is a constraint to crop production, there is little information on the spatial extent of spread and infestation severity of the weed in some SSA countries like Malawi and Zambia. This study aimed to use remotely sensed vegetation phenological (n = 11), climatic (n = 3), and soil (n = 4) variables to develop a data-driven ecological niche model to estimate Striga (Striga asiatica) spatial distribution patterns over Malawi and Zambia, respectively. Vegetation phenological variables were calculated from 250-m enhanced vegetation index (EVI) timeline data, spanning 2013 to 2016. A multicollinearity test was performed on all 18 predictor variables using the variance inflation factor (VIF) and Pearson's correlation approach. From the initial 18 variables, 12 non-correlated predictor variables were selected to predict Striga risk zones over the two focus countries. The variable "start of the season" (start of the rainy season) showed the highest model relevance, contributing 26.8% and 37.9% to Striga risk models for Malawi and Zambia, respectively. This indicates that the crop planting date influences the occurrence and the level of Striga infestation. The resultant occurrence maps revealed interesting spatial patterns; while a very high Striga occurrence was predicted for central Malawi and eastern Zambia (mono-cultural maize growing areas), lower occurrence rates were found in the northern regions. Our study shows the possibilities of integrating various ecological factors with a better spatial and temporal resolution for operational and explicit monitoring of Striga-affected areas in SSA. The explicit identification of Striga "hotspot" areas is crucial for effectively informing intervention activities on the ground.
Keywords: Ecological niche modeling; Food security; Invasive weeds; Phenology; Remote sensing; Sub-Saharan Africa.
© 2022. The Author(s), under exclusive licence to Springer Nature Switzerland AG.