This study compared unsupervised and supervised machine learning, and deep learning (U-Net) classifiers on Unmanned Aerial Vehicle (UAV) multispectral imagery to identify nitrogen status in potato crops under nitrogen (N) fertilization treatments, irrigation (I), and their interaction (N × I). The U-Net model outperformed all other methods, achieving accuracies for crop nitrogen status of 65–99% in N, 84–100% in I, and 41–82% in N × I treatments, with variation due to different input data. Supervised machine learning also performed well, with Support Vector Machine achieving 53–87, 66–86, and 32–66% respectively, and Random Forest 61–96, 70–81, and 33–65%. Unsupervised K-means yielded the lowest accuracies (47–58, 9–65, and 8–34%), demonstrating necessity of substantial supervision to delineate crop nitrogen and water status. These findings were confirmed by repeated analyses of UAV imagery acquired later in the growing season with consistent results. Comparable classification performance was observed for crop water status and leaf area index at both time points. Despite being demonstrated in a single-field, single-crop framework, the results provide proof of concept for applying deep learning classifiers to detect subtle nitrogen and water stress under field conditions in precision agriculture. Future research could test diverse agroecosystems and growing seasons, alternative deep learning algorithms, and sensor data fusion to improve classification accuracies.

Crop failures are common in rain-fed farming in sub-Saharan Africa, especially in water-scarce South Africa. Inadequate rainfall necessitates innovative solutions to enhance food production. Water-saving irrigation technologies can significantly reduce crop failures, particularly for smallholder farms with limited access to irrigation water. This study evaluated the effects of Self-Regulating, Low-Energy, Clay-Based Irrigation System (SLECI), subsurface (SDI) and surface drip (DI) on the performance of moringa (Moringa oleifera) and cowpea (Vigna unguiculata), cultivated either as mono (sole) crops or in intercropping systems, in an open experimental field in South Africa. The experimental design was a factorial Randomized Complete Block Design (RCBD) replicated three times. The main aim was to assess water productivity and yield performance in different irrigation systems over two growing seasons. The results showed that the SLECI irrigation system was more suitable for M. oleifera, while V. unguiculata performed best with standard drip irrigation. Moringa oleifera fresh leaf biomass was higher under SLECI with sand around the clay element and surface drip irrigation with 1.42 t/ha, followed by the SLECI treatment without sand with 1.25 t/ha, while the least yield was noted in subsurface drip irrigation treatment with 1.18 t/ha. Vigna unguiculata (a dual-purpose crop for grain and leaves) produced higher total fresh biomass yield under subsurface drip irrigation treatment with 66.26 t/ha, followed by the SLECI treatment without sand (61.51 t/ha), while drip and SLECI with sand showed similar yield with 52.34 and 52.31 t/ha, respectively. In M. oleifera, the irrigation water productivity (IWP) varied from 0.26 kg/m³ below the surface to 0.65 kg/m³ after the SLECI treatment with sand. IWP in V. unguiculata treatments ranged from 27.52 kg/m³ in SLECI without sand to 9.52 kg/m³ under surface drip irrigation. In addition, chlorophyll content and stem diameter were elevated under SLECI, reflecting enhanced nutrient and water availability. The findings have important implications for sustainable agriculture under water-limited conditions.

Arbuscular mycorrhizal fungi (AMF) and biochar may improve crop growth in drought-prone sandy soils, but their interactive effects under different water regimes remain unclear. We investigated their effects on potato growth and nutrient uptake under water deficit. A semi-field rainout-shelter experiment was conducted in Denmark using low-phosphorus (P) sandy soil (Olsen P: 18.3 mg kg⁻¹), unamended (BC−) or amended with wheat-straw biochar (BC+). Potato plants were grown without (M−) or with Rhizophagus irregularis inoculation (M+) and irrigated at three soil water thresholds (75, 50, and 25% of field capacity). Biochar increased P and nitrogen (N) uptake and tuber biomass, with biomass gains strongly correlated with nutrient uptake. Under high irrigation, BC+M+ increased tuber biomass, P uptake, and N uptake by 55, 73, and 32%, respectively, compared with BC−M−. Without biochar, AMF did not increase nutrient uptake and reduced tuber biomass. AMF root colonization and its soil lipid biomarker were affected by water deficit, with responses also associated with soil pH and P availability. Overall, adequate water availability was critical for realizing the benefits of combined biochar and AMF application in low-P sandy soil.

The high spatio-temporal resolution of UAV sensors requires robust analytical tools to classify subtle agroecosystem variations. This study compared unsupervised, supervised machine learning (ML), and deep learning (U‑Net) classifiers to identify nitrogen (N) and water (I) status, and their interaction (N×I) in potato crops using UAV multispectral imagery. The U‑Net model outperformed all other methods, achieving accuracies of 85% (N), 93% (I), and 70% (N×I). Supervised ML classifiers also performed well and Support Vector Machine achieved 71, 62, and 40% respectively, whereas Random Forest achieved 67, 61, and 40%. The unsupervised K‑means classifier yielded the lowest accuracies (41, 36, and 23%), demonstrating the necessity of substantial supervision to delineate crop N and water properties. These results were confirmed by repeated analysis on UAV imagery acquired later in the season. Deep learning classifiers should be adopted more widely in precision agriculture, as they offer new potential for optimizing N and irrigation co-management under field conditions with subtle spatial variation that is otherwise difficult to capture. Future research should test alternative deep learning algorithms and sensor data fusion to further improve classification accuracies.