Open Access

ARTICLE

Terrain Controls on NDVI Spatial Variability under Post-Harvest Conditions: A UAV-Based Geomorphometric and Machine Learning Approach in Mediterranean Croplands

Jesús Rodrigo-Comino^1,*, María Teresa González-Moreno¹, Lucía Moreno-Cuenca¹, Laura Cambronero-Ruiz¹, Clemente Irigaray², Francisco Serrano Bernardo³, Víctor Hugo Durán Zuazo⁴, Jesús Fernández-Gálvez⁵, Andrés Caballero-Calvo⁵, Víctor Rodríguez-Galiano⁶

1 Departamento de Análisis Geográfico Regional y Geografía Física, Andalusian Research Institute in Data Science and Computational Intelligence, DaSCI, University of Granada, Granada, Spain
2 Department of Civil Engineering, ETSICCP, University of Granada, Campus Fuentenueva s/n, Granada, Spain
3 Department of Civil Engineering, Environmental Technologies Area, University of Granada, Granada, Spain
4 IFAPA Centro “Camino de Purchil”, Camino de Purchil s/n, Granada, Spain
5 Departamento de Análisis Geográfico Regional y Geografía Física, Facultad de Filosofía y Letras, Campus Universitario de Cartuja, Universidad de Granada, Granada, Spain
6 Departamento de Geografía Física y Análisis Geográfico Regional, Universidad de Sevilla, Seville, Spain

* Corresponding Author: Jesús Rodrigo-Comino. Email:

(This article belongs to the Special Issue: Resource and Environmental Information Modeling - 2nd Edition)

Revue Internationale de Géomatique 2026, 35, 333-349. https://doi.org/10.32604/rig.2026.081503

Received 03 March 2026; Accepted 20 May 2026; Issue published 11 June 2026

Abstract

Soil degradation in Mediterranean agricultural systems is strongly conditioned by topography, water redistribution and solar exposure, factors that can be effectively studied using very high-resolution remote sensing. This study evaluates the potential of Unmanned Aerial Vehicle (UAV)-derived geomorphometry combined with machine learning techniques to analyse the spatial variability of the Normalized Difference Vegetation Index (NDVI) as a surface spectral response under post-harvest conditions in a Mediterranean cereal field affected by soil degradation and gully erosion, located near Casabermeja (Málaga, southern Spain). High-resolution RGB and multispectral UAV data were used to generate a Digital Terrain Model (DTM), multi-scale local relief metrics, hydrological indices and curvature derivatives, together with NDVI maps acquired after harvest under dry and compacted soil conditions. A Random Forest regression model was applied using 3000 sampling points to link geomorphometric variables with NDVI spatial patterns. Although predictive performance was moderate (R² = 0.31; RMSE = 0.04), variable-importance analysis identified the main terrain-related factors associated with spatial variability in the spectral signal, highlighting slope, solar exposure and erosion- and flow-convergence-related indices. The results demonstrate the usefulness of UAV-based geomorphometry and machine learning as diagnostic tools for analysing terrain-controlled surface spectral patterns and identifying areas potentially affected by soil degradation processes in Mediterranean agroecosystems.

---

Keywords

---