Improved food-insecurity prediction in smallholder-dominated landscapes using MODIS Enhanced Vegetation Index and Google Earth Engine: a case study in South Central Ethiopia
Recent droughts and food insecurity underline the need for objective, timely, spatially explicit food aid prediction in Ethiopia. We developed a generic user-friendly method to detect greening of agricultural areas and derive predictions of agricultural production for potentially food-insecure areas...
Guardado en:
| Autores principales: | , , |
|---|---|
| Formato: | article |
| Lenguaje: | EN |
| Publicado: |
Taylor & Francis Group
2021
|
| Materias: | |
| Acceso en línea: | https://doaj.org/article/05cbc85eca0e40e681f8a75a568b1b44 |
| Etiquetas: |
Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
|
| Sumario: | Recent droughts and food insecurity underline the need for objective, timely, spatially explicit food aid prediction in Ethiopia. We developed a generic user-friendly method to detect greening of agricultural areas and derive predictions of agricultural production for potentially food-insecure areas. We used the Enhanced Vegetation Index (EVI) from combined Terra/Aqua MODIS (Moderate Resolution Imaging Spectroradiometer) images to generate EVI time series over multiple growing seasons. Maximum seasonal greening (EVImax), as proxy for biomass and expected crop yield, was related to rainfall variability and to indicate areas of risk for crop failure due to drought within the necessary reaction time for emergency aid. Four agroecological zones were covered from 2003 to 2019. Vegetation periods per 250m pixel were calculated back from EVImax. EVImax was validated against measured yields on large-scale farms. Interannual means and variability of EVImax served to assess production and drought risk. Yield predictions corresponded well with wheat production (r2≅0.5 p≤0.05). High temporal variability and low absolute EVI indicated drought-prone areas. EVI was positively correlated with rainfall data in cropped drought-prone areas (r2≅0.4, p≤0.05), but negatively in temporally water-logged highlands (r2≅0.3, p≤0.05). Our user-friendly approach on Google Earth Engine can accurately detect imminent food insecurity and facilitate timely interventions. |
|---|