Moroccan installations of solar photovoltaic panels experience operational difficulties due to shading and vegetation-related soiling, which reduce energy output by 15–30%. Most monitoring systems depend upon a single vegetation index, which can reduce the accuracy of detecting even moderate anomalies. This paper presents a novel integration of multi-criteria hierarchical labeling with dual-branch deep learning for enhanced vegetation anomaly detection. We combined MODIS (2000–2015) and Sentinel-2 (2015–2025) images and NASA POWER weather records to study a 25-year vegetation record using multi-source satellite data in 5 of Morocco’s ecologically diverse zones. We introduced a three-class hierarchical labeling scheme (normal, moderate, severe) for dynamic vegetation models based on combined vegetation indices (NDVI, EVI, NDWI) and meteorological thresholds. The proposed dual-branch architecture uses independent data streams for unfused data, which include temporal multi-scale CNNs (TMSCNN) for spatiotemporal modeling and bidirectional LSTMs for weather-integrated vegetation data. Systematic ablation studies show improvements from using NDVI (68.98%) to multispectral indices (77.74%), meteorological integration (81.02%), and a final accuracy of 82.34% ± 0.88%. The moderate anomaly class exhibits lower precision (65%), demonstrating the challenge of operationalizing severity-based anomaly classification. This work integrates hierarchical, multi-criteria labeling and hybrid deep learning for solar photovoltaic vegetation monitoring.

Climate change is recognized as a global threat, with arid and semi-arid regions highly vulnerable due to fragile ecosystems and limited water. This study analyzes long-term climate trends in Errachidia, Morocco, using high-resolution hourly meteorological data from 2010 to 2024. By applying both time series decomposition and Long Short-Term Memory (LSTM) neural networks, we did identify clear patterns of warming and climatic variability, providing localized insights that support climate adaptation strategies in line with Sustainable Development Goal 13. The results reveal a mean annual temperature increase of 0.1142°C, amounting to approximately 1.6°C over the 14-year period, while seasonal decomposition highlights particularly intense warming during autumn (+0.1666°C per year), with summer temperatures peaking around 35°C between 2022 and 2024. The study showed a frequency of extreme heat events (defined by the 95th percentile) nearing 500 occurrences annually, while the number of cold days has significantly declined since 2020, decreasing by an average of 9.29 events per year. Results also showed wind direction has shifted notably toward the north (from 25.7% to 30.4% occurrence between 2011 and 2023), although wind speed remained largely stable. In this study, we also used LSTM modeling, enhanced with proper inverse scaling and evaluation, which showed improved predictive performance. The temperature model achieved an RMSE of 0.43°C and an MAE of 0.32°C, confirming observed trends and highlighting challenges in modeling extremes in arid environments. The wind speed model showed lower precision due to inherent volatility, but still captured key directional trends. This work provides data-driven climate insights for an underrepresented North African region. Its findings have implications for agriculture, water management, and public health, offering a reproducible framework for future research. By integrating meteorological analysis with artificial intelligence, it highlights the need for targeted climate adaptation in Errachidia and other vulnerable areas across the Global South. Received: 3 April 2025 | Revised: 26 August 2025 | Accepted: 3 December 2025 Conflicts of Interest Azidine Guezzaz is an Editorial Board Member for Journal of Data Science and Intelligent Systems and was not involved in the editorial review or the decision to publish this article. The authors declare that they have no conflicts of interest to this work. Data Availability Statement The raw meteorological observations used in this study originate from publicly available Aeronautical Meteorological Messages for Errachidia Airport, which are routinely disseminated by official meteorological and aviation weather services. The raw Aeronautical Meteorological Messages data are openly accessible, for example, via the NOAA Aviation Weather Center: https://aviationweather.gov/data/metar/?ids=GMFK. No restrictions apply to the use of the raw data. The processed dataset generated during this study is available from the corresponding author upon request. Author Contribution Statement Anas Kabbori: Conceptualization, Methodology, Software, Validation, Formal analysis, Resources, Data curation, Writing — original draft, Writing — review & editing, Visualization. Chahrazad Zargane: Conceptualization, Validation, Formal analysis, Investigation, Resources, Writing — original draft, Writing — review & editing. Azidine Guezzaz: Methodology, Software, Validation, Resources, Data curation, Writing — original draft, Writing — review & editing, Supervision, Project administration. Said Benkirane: Methodology, Validation, Investigation, Resources, Writing — original draft, Writing — review & editing, Supervision. Mourade Azrour: Software, Validation, Investigation, Resources, Writing — original draft, Writing — review & editing, Visualization, Supervision.