Spatio-Temporal Maize Yield Prediction Using Sparse Input Data for Mali with Craft

52 Pages Posted: 11 Jun 2024

See all articles by Steven Ndung'u

Steven Ndung'u

affiliation not provided to SSRN

Pierre C. Sibiry Traore

affiliation not provided to SSRN

Vakhtang Shelia

University of Florida

Andree Nenkam Mentho

affiliation not provided to SSRN

Janet Mumo Mutuku

affiliation not provided to SSRN

Sridhar Gummadi

International Center for Biosaline Agriculture

James Hansen

Columbia University

Anthony Whitbread

affiliation not provided to SSRN

Gerrit Hoogenboom

University of Florida

Abstract

Significant progress has been made in the application of mechanistic crop models to define new and enhanced strategies for improved risk management, efficient crop production, sustainable cropping systems, and assessment of the impact of climate change and variability on agricultural systems. The West African region continues to experience shifting climatic patterns and increased land degradation with far-reaching consequences on the production of small-holder farmers. The goal of this study was to assess maize cropping systems gridded simulations' applicability using the CCAFS Regional Forecasting Toolbox (CRAFT) considering recent climate variability in the Sudano-Sahelian region of West Africa using available gridded datasets. CRAFT is a flexible and adaptable software designed to produce multiple simulation scenarios, maps and other interactive visualizations with crop models based on spatially distributed weather conditions, soil properties and management practices. In this study, CRAFT was used to perform hindcast analyses in the mandate area of the Compagnie Malienne de Développement des Textiles (CMDT), Southern Mali, focusing on maize production over 26 years from 1990 to 2015. Among the considered weather sources for the study, satellite gridded datasets from the Enhancing National Climate Services (ENACTS) initiative showed low systematic errors and bias. CRAFT generated estimates for 2010 - 2015 maize yield aligned closely with measured data from the CMDT operational monitoring system. Additionally, the relative mean absolute error (RMAE) and relative root mean square error (RRMSE) metrics of 75% of the sectors in the CMDT region were below 35%. This demonstrates that CRAFT offers new insights into regional yield predictions for maize and provides an opportunity to objectively improve food security predictions in the context of rapid socio-economic and environmental changes.

Keywords: Climate risk, Smallholder agriculture, Gridded simulations, Edaphic factors, Decision support, Food security

Suggested Citation

Ndung'u, Steven and Traore, Pierre C. Sibiry and Shelia, Vakhtang and Nenkam Mentho, Andree and Mutuku, Janet Mumo and Gummadi, Sridhar and Hansen, James and Whitbread, Anthony and Hoogenboom, Gerrit, Spatio-Temporal Maize Yield Prediction Using Sparse Input Data for Mali with Craft. Available at SSRN: https://ssrn.com/abstract=4860735 or http://dx.doi.org/10.2139/ssrn.4860735

Steven Ndung'u (Contact Author)

affiliation not provided to SSRN ( email )

No Address Available

Pierre C. Sibiry Traore

affiliation not provided to SSRN ( email )

No Address Available

Vakhtang Shelia

University of Florida ( email )

Andree Nenkam Mentho

affiliation not provided to SSRN ( email )

No Address Available

Janet Mumo Mutuku

affiliation not provided to SSRN ( email )

No Address Available

Sridhar Gummadi

International Center for Biosaline Agriculture ( email )

P.O. Box 14660
Dubai
United Arab Emirates

James Hansen

Columbia University ( email )

Anthony Whitbread

affiliation not provided to SSRN ( email )

No Address Available

Gerrit Hoogenboom

University of Florida ( email )

Do you have a job opening that you would like to promote on SSRN?

Paper statistics

Downloads
7
Abstract Views
56
PlumX Metrics