From forest to farmland: Tracking time series variations in soil quality in semiarid oak forest

• Published in: Geoderma Regional Vol. 42; p. e00974
• Main Authors: Fathy, Hojat, Heydari, Mehdi, Fathizad, Hassan, Hosseinzadeh, Jaafar, Najafifar, Ali, Mousavi, Seyed Roohollah, Jafarzadeh, Aliakbar, Heung, Brandon
• Format: Journal Article
• Language: English
• Published: Elsevier B.V 01.09.2025
• Subjects: Environmental variables; Land use changes; Machine learning; Soil quality assessment; Temporal data; Soil quality assessment; Land use changes; Temporal data; Environmental variables; Machine learning
• DOI: 10.1016/j.geodrs.2025.e00974
• ISSN: 2352-0094

Assessing soil quality (SQ) is essential for sustainable development, as soils play a fundamental role in ecosystems, influencing habitat characteristics, ecosystem stability, and biogeochemical cycles. This study examines the spatio-temporal variations of the Soil Quality Index (SQI) in response to land use changes within a semi-arid oak forest ecosystem in western Iran from 1985 to 2021. To conduct the research, 150 surface soil samples (0–30 cm) were collected using the conditional Latin hypercube sampling method. The physical and chemical attributes of these samples were analyzed to determine the SQI using the Integrated Weighted Index method. Additionally, various environmental variables—including geological maps, topographic attributes, and proximity to streams and roads—were incorporated into the analysis. Time series maps of vegetation indices and land use were generated using Landsat 8 imagery (covering 1985, 2000, 2010, and 2021) to support SQI spatial prediction. The Random Forest (RF) model was employed to model and map both SQI and land use changes over time. The RF model demonstrated high accuracy, achieving a Kappa index of 92.3 %–96.0 % for land use predictions and an R2 value of 0.80–0.88 for SQI predictions. Spatial modeling results identified four key environmental factors influencing SQI: convergence index, vector terrain ruggedness, distance from streams, and minimum noise fraction of Landsat bands. Analysis of SQ changes from 1985 to 2021 revealed a significant shift in the distribution of SQI classes across various land uses. The proportion of low-SQ areas increased by 20.3 % in forests, 22.6 % in agricultural lands, and 29.7 % in rangelands, indicating widespread soil degradation. Notably, in the central part of the study area—where rangeland and agricultural lands have expanded—SQI exhibited a downward trend, further reflecting soil deterioration associated with these land uses. This research highlights the critical need for ongoing SQ monitoring and sustainable land management measures to mitigate soil degradation in semi-arid ecosystems. The findings emphasize the detrimental impact of land use changes on soil health, particularly the conversion of forests to agricultural lands, which has contributed to severe soil degradation and biodiversity loss. Implementing conservation strategies, such as reforestation and improved agricultural practices, is essential to preserving soil quality in the region.