Published November 30, 2022 | Version v1
DE-AFRICA Knowledge Package Open

Forecasting Cropland Vegetation Condition Using Digital Earth Africa Data Cube

Description

 

What is it about

Background

This notebook conducts time-series forecasting of vegetation condition (NDVI) using SARIMAX, a variation on autoregressive-moving-average (ARMA) models which includes an integrated (I) component to difference the time-series so it becomes stationary, a seasonal (S) component, and can consider exogenous (X) variables. In this example, we will conduct a forecast on a univariate NDVI time series. That is our forecast built on temporal patterns in NDVI. Conversely, multivariate approaches can account for influences of variables such as soil moisture and rainfall.

This notebook allows for easy replication of analysis across Africa, by simply changing the analysis parameters. The user guide provides an overview of the research, resources used/supplied, and directions on reproducing the research.

This knowledge package includes the necessary data, computational resources, and instructions to reproduce the methodology used in Digital Earth Africa's 'Forecasting cropland vegetation condition' Jupyter Notebook. 

The User's Guide provides an overview of the application.

Knowledge Resources

Files

File view

Additional details

See also

Created:
January 4, 2023
Modified:
January 2, 2025