Orhan Dengiz

The rapid and continuous growth of the human population since the Industrial Revolution has posed a significant threat to the quantity and quality of our water, air and soil resources, and has weak...

Soil quality is a key indicator of sustainable land management because it reflects the capacity of soil to perform essential physical, chemical, and biological functions. This study aimed to evaluate the spatial variability of soil quality under different land use and land cover types within a micro-catchment located in Bafra District, Samsun Province, in the Central Black Sea Region of Türkiye. The study area includes forest, pasture, hazelnut orchards, and cultivated agricultural lands under semi-humid ecological conditions. A total of 27 soil indicators representing physical, chemical, biological, and fertility-related properties were used to assess soil quality. Surface soil samples (0–30 cm) were collected from 62 sampling points distributed across the study area. Soil quality indicators were grouped under four main criteria and weighted using the Analytic Hierarchy Process (AHP). Linear and non-linear Standard Scoring Functions (SSF) were then applied to normalize the indicator values and calculate linear and non-linear soil quality indices (L-SQI and NL-SQI). Spatial distribution maps of both indices were produced in a GIS environment using geostatistical interpolation methods. The results showed that the spatial patterns of L-SQI and NL-SQI were highly similar across the study area. Low soil quality values were mainly concentrated in the eastern parts of the micro-catchment, whereas relatively higher soil quality values were observed in the northern and southern parts. Forested areas generally exhibited higher soil quality, while hazelnut-dominated areas were associated with relatively lower soil quality values. The findings demonstrate that integrating AHP, SSF, geostatistics, and GIS provides an effective framework for evaluating and mapping soil quality under different land use systems and can support sustainable land management decisions in regions with similar ecological conditions.

Toprak erozyonu, doğal kaynakların sürdürülebilir yönetimini tehdit eden en önemli arazi bozunumu süreçlerinden biridir. Bu çalışmada, Rize ili sınırları içerisinde yer alan bir mikro havzada erozyon riskinin mekânsal dağılımının belirlenmesi amacıyla ICONA erozyon risk modeli kullanılmıştır. Çalışma kapsamında, arazi örtüsü, arazi kullanımı-arazi örtüsü, jeoloji ve eğim gibi erozyon süreçlerini etkileyen temel çevresel faktörler Coğrafi Bilgi Sistemleri (CBS) ortamında analiz edilmiştir. Öncelikle arazi örtüsü ve arazi kullanım haritaları birleştirilerek toprak koruma haritası oluşturulmuş, ardından jeoloji ve eğim katmanlarının entegrasyonu ile potansiyel erozyon risk haritası elde edilmiştir. Son aşamada ise bu iki katmanın çakıştırılmasıyla ICONA erozyon risk haritası hazırlanmıştır. Elde edilen sonuçlar, çalışma alanının önemli bir bölümünde düşük ve orta düzeyde erozyon riskinin hâkim olduğunu göstermiştir. Bununla birlikte, özellikle havzanın güney kesimlerinde eğim değerlerinin yüksek olması, litolojik yapının ayrışmaya daha duyarlı özellikler göstermesi ve bitki örtüsünün görece zayıf olması nedeniyle yüksek ve çok yüksek erozyon riskine sahip alanların yoğunlaştığı belirlenmiştir. Çalışma sonuçları, erozyon riskinin havza ölçeğinde mekânsal olarak belirlenmesinde ICONA modelinin etkin bir araç olduğunu ortaya koymaktadır. Ayrıca elde edilen bulguların, erozyon kontrolü ve arazi yönetimi planlamalarında öncelikli alanların belirlenmesine katkı sağlayabileceği düşünülmektedir.

Saline soils represent a major challenge to agricultural productivity, particularly in regions like Belozem, Bulgaria, where salinisation is both naturally occurring and anthropogenically induced. This study investigated the microbial status of strongly saline soils through a combination of conventional microbiological methods and metagenomic sequencing. Soil respiration, substrate induced respiration, colonyforming units, and enzymatic activities (β-glucosidase and dehydrogenase) were assessed, alongside high-throughput 16S rRNA gene sequencing. The results indicated that increased soil salinity and gravimetric water content negatively affected microbial respiration and diversity. The microbial community was dominated by halotolerant taxa including Actinobacteriota, Proteobacteria, and Firmicutes. Soil respiration was significantly correlated with moisture, and substrate induced respiration values revealed high microbial dormancy under saline stress. Enzyme assays indicated suppressed metabolic activity in high-salt environments, particularly in soils with the lowest pH and highest conductivity values. Metagenomic analysis revealed variations in alpha and beta diversity across the three soil types, reflecting salinity-induced shifts in microbial community structure and function. These findings highlight the ecological consequences of salinity on soil microbial dynamics and suggest that metagenomic approaches can offer valuable insights for managing saline-affected ecosystems.

Soil erodibility (USLE‐K) is a critical indicator of soil susceptibility to erosion; however, its accurate estimation remains challenging due to complex and nonlinear interactions among soil physicochemical properties. Most existing studies focus either on traditional machine learning techniques or on individual deep learning models, often addressing classification or regression tasks separately and rarely integrating spatial validation of predictive results. The present study addresses these limitations by proposing an integrated comparative framework that simultaneously evaluates machine learning and advanced deep learning algorithms for both classification and continuous prediction of soil erodibility. Random Forest (RF) was compared with deep learning architectures including 1D CNN, CNN + LSTM, TabNet, RNN, LSTM, and GRU using a large dataset derived from field measurements and laboratory soil analyses in the Sakarya Basin. Model performance was assessed using standard classification and regression metrics, while predictive outputs were further evaluated through geostatistical interpolation to examine spatial consistency. The TabNet classifier achieved the highest classification accuracy (94%), demonstrating its effectiveness in capturing complex feature interactions in tabular soil data. For regression‐based prediction, the GRU model exhibited superior performance with an R ² value of 94%, outperforming both other deep learning models and the classical RF approach. Statistical analyses indicated that regression‐based deep learning models produced largely similar predictions, a finding supported by strong agreement in spatial distribution maps. The added value of this study lies in its combined evaluation of predictive accuracy and spatial coherence, demonstrating that deep learning models—particularly TabNet and GRU—provide reliable and spatially consistent soil erodibility assessments. The proposed framework offers a practical decision‐support tool for soil erosion risk mapping, land‐use planning, and the development of effective soil conservation and erosion control strategies in heterogeneous environmental systems.