Publications

Fuentes, I; Al-Shammari, D; Al-Nasrawi, AKM; Wang, Y; Wang, J; Lebrini, Y; Chen, Y; Jones, BG; Bishop, TFA (2025). The normalised difference vegetation index as an analytic tool for wheat crop yield prediction: A review and meta-analysis. PRECISION AGRICULTURE, 26(4), 55.

Abstract
The normalised difference vegetation index (NDVI) is widely used for crop yield prediction. Several studies have shown that there is a positive correlation between NDVI and crop yield, with higher NDVI values indicating healthier and more productive crops. However, various factors can influence the accuracy of the NDVI-crop yield relationship. A systematic review, meta-analysis, and topic modelling analysis were conducted to summarise and quantify the existing evidence of this relationship. More specifically, this review evaluated studies that used NDVI as a tool for wheat crop yield prediction and applied the Latent Dirichlet Allocation (LDA) model to uncover the thematic structure of existing literature. Results show that while NDVI can serve as a standalone predictor, its generalisability and accuracy are limited by factors like observation timing and the chosen statistical approach. Notably, NDVI saturation, particularly above values of 0.75, leads to inaccurate yield estimations, highlighting the need for caution when using peak NDVI. Spatial, temporal, spectral, and radiometric uncertainties may further introduce errors that impact yield predictions. Additionally, agronomic and environmental conditions significantly influence the NDVI-yield relationship, emphasising the complexity of yield estimation models. The meta-analysis revealed substantial variation among studies due to the source of NDVI and the sampling statistic used. Therefore, relying solely on NDVI for crop yield prediction in simple linear regression models can lead to unrealistic yield estimations, especially if large scales and peak NDVI are utilised. Complementing these findings, the LDA analysis identified five key topics, with Vegetation modelling emerging as the dominant theme, reflecting NDVI's central role in crop yield prediction. The co-occurrence of topics such as Vegetation modelling and Water, soil, and productivity highlights the interconnected nature of research on NDVI and its integration with studies on environmental and agricultural factors. This reinforces the need to consider multifaceted drivers influencing the NDVI-yield relationship to enhance the accuracy and applicability of crop yield predictions.

DOI:
10.1007/s11119-025-10247-z

ISSN:
1573-1618