Effects of tree-stripes on crop growth in agroforestry systems using unmanned aerial systems-based analysis

Abstract

Abstract Agroforestry systems (AFS) offer a promising strategy to address environmental challenges while supporting rising food demands. However, the complex interactions between trees and crops complicate research, particularly regarding their effects on crop yields. This study presents a methodological approach using multispectral unmanned aerial system (UAS) data to investigate a maize-cultivated alley cropping system in eastern Germany as a case study. Growth parameters, namely the Normalized Difference Vegetation Index (NDVI) and plant height, were derived as proxies for yield and analyzed in relation to the distance from tree stripes. Additionally, direction-dependent regression analyses were conducted to assess whether spatial variations in the field could be attributed to the trees. Two distinct patterns emerged: first, a pronounced increase in NDVI was observed at close proximity to the trees, correlated with tree height and schematically illustrated for two representative tree stripes; second, at greater distances, fluctuations in NDVI were associated with the trees but lacked consistent directional trends. Considerable inconsistencies were also observed in plant height variations. The discussion highlights potential drivers of the close-range NDVI increase, the applicability of UAS for AFS research, and limitations in generalizing findings from a single case study. Overall, the results demonstrate that tree effects on crop growth and vitality are detectable but marginal in terms of their influence on maize yields at this site, while showcasing the utility of UAS-based approaches for field-scale analysis of AFS.

Plant phenotyping relevance

マルチスペクトルUASからNDVIと植物高を抽出し、樹木からの距離に伴う作物形質を解析する手法の実質的適用が研究の中心であり、単なるルーチン測定を超える。

Code and data availability

The paper's UAS-derived NDVI/plant-height datasets and analysis are not publicly deposited; the Data availability statement says they are available from the corresponding author on reasonable request. All URLs in the reference list are cited third-party sources or software, not paper-specific assets.

No evidence-backed public reproduction asset is currently recorded.