Geoderma最新文献

{"title":"Quantitative evaluation of carbon dioxide emissions from the subsoils of volcanic and non-volcanic ash soils in temperate forest ecosystems","authors":"Yukiko Abe ,&nbsp;Masataka Nakayama ,&nbsp;Mariko Atarashi-Andoh ,&nbsp;Takeshi Tange ,&nbsp;Haruo Sawada ,&nbsp;Naishen Liang ,&nbsp;Jun Koarashi","doi":"10.1016/j.geoderma.2025.117221","DOIUrl":"10.1016/j.geoderma.2025.117221","url":null,"abstract":"<div><div>Subsoils (typically below a depth of 30 cm) contain more than half of global soil carbon (C) as soil organic C (SOC). However, the extent to which subsoil SOC contributes to the global C cycle and the factors that control it are unclear because quantitative evaluation of carbon dioxide (CO₂) emission from subsoils through direct observations is limited. This study aimed to quantify CO₂ emission from subsoils and determine factors that control CO₂ emission, focusing on the decomposability of soil organic matter (SOM) and the characteristics of the mineral–SOM association in soils. Therefore, a laboratory incubation experiment was conducted using surface soils (0–10 cm and 10–25 cm depth) and subsoils (30–45 cm and 45–60 cm depth) collected from four Japanese forest sites with two different soil types (volcanic ash and non-volcanic ash soils). The CO₂ emission from the subsoils was found to be responsible for 6 %–23 % of total CO₂ emission from the upper 60-cm mineral soil across all sites. Radiocarbon signatures of CO₂ released from the subsoils indicated the decomposition of decades-old SOM in the subsoils. The correlations between CO₂ emission rate and soil factors across both soil types suggested that the CO₂ emission from the subsoils is mainly controlled by the amounts of SOC easily available to soil microbes and microbial biomass C, not by the amounts of reactive minerals. Given the potential active participation of subsoils in terrestrial C cycling, most of the current soil C models that ignore subsoil C cycling are likely to underestimate the response of soil C to future climate change. The quantitative and mechanistic understanding of C cycling through a huge subsoil C pool is critical to accurately evaluating the role of soil C in the global C balance.</div></div>","PeriodicalId":12511,"journal":{"name":"Geoderma","volume":"455 ","pages":"Article 117221"},"PeriodicalIF":5.6,"publicationDate":"2025-02-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143444368","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"农林科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Effects of tree fall on soil Collembola: Disentangling the role of gap formation and deadwood addition","authors":"Yan Zhang ,&nbsp;Zhou Zheng ,&nbsp;André Junggebauer ,&nbsp;Melanie M. Pollierer ,&nbsp;Stefan Scheu","doi":"10.1016/j.geoderma.2025.117217","DOIUrl":"10.1016/j.geoderma.2025.117217","url":null,"abstract":"<div><div>Increased frequency of climate extremes causes large scale forest decline associated with gap formation and input of deadwood to the forest floor, which largely changes soil systems. However, for disentangling the effects of gap formation and deadwood addition, experimental manipulations allowing to separate the effects of each are needed. Based on a large-scale full-factorial forest gap and deadwood addition experiment, we analyzed the response of soil invertebrates (Collembola) to gap formation and deadwood addition across three geographical regions in Germany i.e., the Alb, Hainich and Schorfheide. Both gap formation and deadwood addition altered the taxonomic richness, density and traits of Collembola communities. The effects of gap formation and deadwood addition were independent of each other but varied among regions, reflecting the importance of both geographic and historical context, as well as environmental changes such as variations in climate. Gap formation strongly decreased total density of Collembola in the Hainich but increased it in the Schorfheide, indicating that the effect is negative in regions with high precipitation and deep soils but positive in regions with low precipitation and shallow soils. Deadwood addition little affected Collembola density but restructured the community composition and increased overall functional and species taxonomic richness, presumably by expanding niche space by increased habitat heterogeneity. Gap formation filtered for small-sized and soil-living species via decreased soil moisture, but did not affect other traits such as number of ocelli and reproduction mode. The results suggest that gap formation and deadwood addition affect Collembola communities and traits in an independent way. Overall, the results indicate that deadwood is pivotal for soil diversity conservation, and forest gaps detrimentally affect animals deeper in soil being adapted to moist conditions.</div></div>","PeriodicalId":12511,"journal":{"name":"Geoderma","volume":"455 ","pages":"Article 117217"},"PeriodicalIF":5.6,"publicationDate":"2025-02-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143444370","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"农林科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Infiltration mechanism and source of soil water in alpine meadows based on stable isotope tracing","authors":"Li Zongjie ,&nbsp;Xu Bin ,&nbsp;Liu Xiaoying ,&nbsp;Li Zongxing ,&nbsp;Feng Qi ,&nbsp;Wang Dongpeng ,&nbsp;Zhang Wenbao ,&nbsp;Li Hao ,&nbsp;Liu Fang","doi":"10.1016/j.geoderma.2025.117224","DOIUrl":"10.1016/j.geoderma.2025.117224","url":null,"abstract":"<div><div>To strengthen the management and protection of soil water resources in the Yellow River source area(SRYR), this study conducted a quantitative analysis of the infiltration of soil water in alpine meadows. Meanwhile, the recharge source and mode of soil water in the SRYR were analyzed. The results indicated that both piston flow mode and priority flow mode were present, with the piston flow mode being predominant. The contributions of priority flow mode to deep soil water in different parts of the SRYR were as follows: South &gt; West &gt; North &gt; East. The contributions of piston flow mode to deep soil water were as follows: East &gt; North &gt; West &gt; South. In the heavy ablation period in 2021, precipitation was identified as the primary recharge source of soil water in the SRYR. The contribution rates of precipitation to soil water decreased with the increase in soil depth. The contribution rate of precipitation to soil water on the sunny slope was slightly higher than that on the shady slope. Different vegetation types had obvious effects on the recharge proportions of soil water. In addition, the contribution rates of precipitation to soil water decreased with the increase in altitude, while the contribution rates of ground ice increased with the increase in altitude. This study can provide theoretical support for soil water management and protection in the SRYR, which is conducive to the sustainable development of soil water resources.</div></div>","PeriodicalId":12511,"journal":{"name":"Geoderma","volume":"455 ","pages":"Article 117224"},"PeriodicalIF":5.6,"publicationDate":"2025-02-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143444371","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"农林科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Evaluation and improvement of spatiotemporal estimation and transferability of multi-layer and profile soil moisture in the Qinghai Lake and Heihe River basins using multi-strategy constraints","authors":"Jiaxin Qian ,&nbsp;Jie Yang ,&nbsp;Weidong Sun ,&nbsp;Lingli Zhao ,&nbsp;Lei Shi ,&nbsp;Hongtao Shi ,&nbsp;Lu Liao ,&nbsp;Chaoya Dang ,&nbsp;Qi Dou","doi":"10.1016/j.geoderma.2025.117222","DOIUrl":"10.1016/j.geoderma.2025.117222","url":null,"abstract":"<div><div>The machine learning regression (MLR) algorithms have brought new insights into soil moisture (SM) estimation. However, few studies have explored the potential of MLR algorithms for multi-layer and profile SM modeling, as well as their spatiotemporal transferability, which are important for practical deployment and application. In this study, the dual-polarization C-band radar data was used as the core to construct a multi-layer and profile SM estimation framework, constrained by multi-source auxiliary data (optical vegetation descriptors, soil properties, and terrain factors). Validation was carried out in two new SM observation networks: the Qinghai Lake basin (QLB-NET) and Heihe River basin (WATERNET). The results shown that the multi-output and multi-input stacking strategy regression (SSR) model demonstrated excellent spatiotemporal extensibility (RMSE = 0.027–0.044 cm³/cm³) and interannual transferability (RMSE = 0.031–0.055 cm³/cm³) in multi-layer and profile SM estimation. However, the cross-spatial transfer accuracy of the SSR model was poor (RMSE &gt; 0.060 cm³/cm³). To address this, two improvement schemes were proposed, focusing on the accessibility of in-situ observation data. The first involved introducing a small number of samples from the target domain to update the hyperparameters in the SSR model. The second method used initial estimates from a scattering model, namely the modified change detection model, to constrain the SSR model and improve cross-spatial transfer accuracy. Both schemes achieved satisfactory transfer accuracy. The former strategy reduced SM estimation errors by 40.8–72.8 % and 24.1–68.1 % across various soil depths for two study areas, while the latter strategy achieved reductions of 30.3–67.2 % and 22.4–68.8 %, respectively. Additionally, factors influencing SM estimation and transfer accuracy were identified, including station difference, SM variability, vegetation cover, soil properties, and imaging orbits. Surprisingly, due to relatively low temporal variability and sensitive to vegetation productivity, the spatiotemporal estimation and transfer accuracy of deeper SM (10–30 cm) was better than that of surface SM (0–10 cm) at most observation stations. The SSR model outperformed deep learning algorithms of different architectures in terms of spatiotemporal estimation and transfer accuracy, operating efficiency, and computational overhead. In conclusion, the framework proposed in this study offers new perspectives and application prospects for remote sensing estimation of multi-layer and profile SM.</div></div>","PeriodicalId":12511,"journal":{"name":"Geoderma","volume":"455 ","pages":"Article 117222"},"PeriodicalIF":5.6,"publicationDate":"2025-02-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143444369","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"农林科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Life history traits in microarthropods: Evidence for a soil animal economics spectrum","authors":"Jing-Zhong Lu ,&nbsp;Tobias Pfingstl ,&nbsp;Robert R. Junker ,&nbsp;Mark Maraun ,&nbsp;Amandine Erktan ,&nbsp;Stefan Scheu","doi":"10.1016/j.geoderma.2025.117206","DOIUrl":"10.1016/j.geoderma.2025.117206","url":null,"abstract":"<div><div>Evolution optimizes the performance of living organisms through budgeting of limited resources, leading to life-history trade-offs. Many life-history traits are related to body size with larger species typically exhibiting a slower pace of life and lower fecundity. However, soil-living organisms may exhibit size-independent life-history strategies due to habitat space constraints, but this has never been tested. Here, we synthesize life-history traits in springtails (Insecta: Collembola) and mites (Acari: Oribatida, Astigmata, Mesostigmata), the most abundant microarthropods worldwide, living mainly in litter and the pore space of soil. We related life-history traits to body size and individual metabolic rate, and showed that life-history traits of soil microarthropods display a trade-off between lifespan and reproductive rate, spanning a continuum from fast to slow life-history strategies. Oribatida exhibit remarkably slow life-histories and long lifespans with lower reproductive rates than Collembola, Astigmata and Mesostigmata. Despite fresh body mass of soil microarthropods varying by three orders of magnitude, fast and slow life-history strategies occurred in all size classes suggesting largely size-independent life-history strategies. Overall, these findings indicate a soil animal economics spectrum that bears key implications for understanding local biodiversity and the coexistence of soil animal species, such as how Collembola and Oribatida coexist worldwide.</div></div>","PeriodicalId":12511,"journal":{"name":"Geoderma","volume":"455 ","pages":"Article 117206"},"PeriodicalIF":5.6,"publicationDate":"2025-02-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143453089","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"农林科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Deep-rooted perennials alter microbial respiration and chemical composition of carbon in density fractions along soil depth profiles","authors":"Kyungjin Min ,&nbsp;Erin Nuccio ,&nbsp;Eric Slessarev ,&nbsp;Megan Kan ,&nbsp;Karis J. McFarlane ,&nbsp;Erik Oerter ,&nbsp;Anna Jurusik ,&nbsp;Gregg Sanford ,&nbsp;Kurt D Thelen ,&nbsp;Jennifer Pett-Ridge ,&nbsp;Asmeret Asefaw Berhe","doi":"10.1016/j.geoderma.2025.117202","DOIUrl":"10.1016/j.geoderma.2025.117202","url":null,"abstract":"<div><div>Growing deep-rooted perennials has been proposed to increase soil organic carbon (SOC) stocks and mitigate CO₂ <!-->emissions. Yet, we know little about the bioavailability and chemical properties of SOC under deep-rooted perennials and shallow-rooted annuals. Improving our understanding of the role of deep-rooted perennials for belowground C storage is critical, as root growth has the potential to both increase SOC stock and accelerate loss of existing SOC. Here, we assessed the effects of &gt;10 years of land conversion from shallow-rooted annuals (maize) to deep-rooted perennials (switchgrass) on SOC bioavailability (microbial respiration, Δ¹⁴C-CO₂), mineral-associated SOC (density fractionation), and SOC turnover and composition (¹⁴C-SOC, DRIFT spectroscopy) in surface soils (0–20 cm) and subsoils (90–120 cm) at two sites with sandy and silty soils. We demonstrate that switchgrass enhanced microbial respiration of recently-fixed C in surface soils. Switchgrass increased Δ¹⁴C values of the free light fraction in subsoil of the sandy site, by supplying aliphatic C (putative simple plant C) into the soil. In contrast, maize input less root C into the soil, and at one site increased the decomposition of older SOC, which indicates that overall microbial C demand outpaced plant C inputs. These results highlight that deep-rooted perennials stimulate the transfer of more atmospheric C to both surface and subsoils than shallow-rooted annuals, that newly generated SOC under deep-rooted perennials is relatively less protected from decomposition, and that reaping the C benefits of deep-rooted perennials could require maintaining the land cover as a perennial cropping system.</div></div>","PeriodicalId":12511,"journal":{"name":"Geoderma","volume":"455 ","pages":"Article 117202"},"PeriodicalIF":5.6,"publicationDate":"2025-02-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143438087","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"农林科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Reevaluating multi-pool first-order kinetic models for fitting soil incubation data","authors":"Shuhao Zhou ,&nbsp;Daifeng Xiang ,&nbsp;Gangsheng Wang ,&nbsp;Liping Zhang ,&nbsp;Zehao Lv ,&nbsp;Shanshan Qi ,&nbsp;Wanyu Li","doi":"10.1016/j.geoderma.2025.117218","DOIUrl":"10.1016/j.geoderma.2025.117218","url":null,"abstract":"<div><div>Soil incubation experiments are frequently conducted to investigate soil carbon (C) cycling and its response to environmental changes. Multi-pool first-order models, which apply constant kinetic rate parameters for each pool, are widely used for fitting these datasets due to their simplicity. However, their ability to accurately represent instantaneous C effluxes is often overlooked, as cumulative effluxes are typically prioritized. Here, we calibrated a three-pool first-order model using a 384-day incubation dataset with fluctuating CO₂ and CH₄ effluxes. Despite constantly good performance for cumulative C effluxes (<em>R</em>² = 0.99–1.00), the <em>R</em>² values vary greatly with experimental conditions for instantaneous effluxes (<em>R</em>² = 0.36–0.93), depending on the frequency of aerobic-anaerobic shifts. The results are regardless of the objective function used for model fitting (i.e., maximizing <em>R</em>² for instantaneous or cumulative C effluxes). Compared to cumulative effluxes, we propose emphasizing instantaneous effluxes and calibrating additional variables in soil C modeling. This strategy can offer more insights for improving predictions and enhancing understanding of soil C-climate feedbacks under dynamic environmental conditions.</div></div>","PeriodicalId":12511,"journal":{"name":"Geoderma","volume":"455 ","pages":"Article 117218"},"PeriodicalIF":5.6,"publicationDate":"2025-02-18","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143430105","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"农林科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Integrated soil health management influences soil properties: Insights from a US Midwest study","authors":"Anuradha Garg ,&nbsp;Samuel Kwakye ,&nbsp;Anna Cates ,&nbsp;Heidi Peterson ,&nbsp;Kathryn LaBine ,&nbsp;Greg Olson ,&nbsp;Vasudha Sharma","doi":"10.1016/j.geoderma.2025.117214","DOIUrl":"10.1016/j.geoderma.2025.117214","url":null,"abstract":"<div><div>This study evaluated the effects of integrated soil health management in the US Upper Midwest over three years (2021–2023), under diverse cropping systems and soil textures. We assessed 15 field pairs, each consisting of one conventional (CV) and one soil health (SH) site, implementing contrasting management. Our analysis focused on four soil organic matter pools, six microbial indicators derived from phospholipid fatty acids (PLFA) and one physical indicator. Log response ratios (LRR) were calculated to compare pair-wise responses between medium and moderately fine-textured soils. Wet aggregate stability (WAS) showed consistent improvement; within each pair, more soil health based principles (reduced tillage, more cover crops and crop diversity) led to greater aggregate stability compared to the paired CV site. Medium-textured soils responded more strongly to soil health management than moderately fine-textured soils. To assess the effects of specific management practices, we built a mixed-effects model with practices and their interactions as fixed effects and soil health indicators as response variables. Results showed that most soil properties were significantly responsive to two management combinations, 1) tillage x cover crops, and 2) tillage x cover crops x crop diversity. Microbial indicators along with potentially mineralizable nitrogen (PMN) exhibited the strongest increases with integrated soil health management (<em>p</em> &lt; 0.05), followed by permanganate oxidizable carbon (POXC) and total N (<em>p</em> &lt; 0.1). Cover cropping alone moderately increased PMN (<em>p</em> &lt; 0.1). While site-specific behavior varied based on texture and management intensities, our overall results supported the adoption of integrated soil health practices for healthier agricultural soils.</div></div>","PeriodicalId":12511,"journal":{"name":"Geoderma","volume":"455 ","pages":"Article 117214"},"PeriodicalIF":5.6,"publicationDate":"2025-02-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143430062","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"农林科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Edaphic and climatic effects on soil water dynamics and infiltration patterns in tropical rainforests","authors":"Xin Zou ,&nbsp;Ankit Shekhar ,&nbsp;Yuxuan Mo ,&nbsp;Ashutosh Kumar Singh ,&nbsp;Xiaojin Jiang ,&nbsp;Wenjie Liu","doi":"10.1016/j.geoderma.2025.117197","DOIUrl":"10.1016/j.geoderma.2025.117197","url":null,"abstract":"<div><div>The preservation of forests’ hydrological functions is pivotal for alleviating land degradation and enhancing ecosystem services. However, understanding of the hydrological processes in tropical rainforests remains inadequate, particularly in the context of a changing climate and complex edaphic conditions. This study aims to quantify the spatiotemporal variations of soil water content and infiltration patterns within tropical rainforests located on distinct slope gradients (flat 0°, moderate 10° and steep 30°), and elucidate the effects of climatic and edaphic factors on these hydrological parameters. Based on the one year observation, it was revealed that the moderate-slope site maintained higher soil water content levels and temporal stability (54.32) compared to the flat (41.85) and steep-slope (42.71) sites across the 0 – 110 cm soil depths. This phenomenon is likely attributed to the superior soil condition at the moderate-slope site, characterized by optimal soil physical properties, a rich content of soil organic matter and abundant root biomass. The regression analysis demonstrated that soil physical properties (represented by capillary holding capacity and bulk density) had significant positive effects on soil water throughout the entire year (R² &gt; 0.6, <em>p</em> &lt; 0.01). The effect of sand content on soil water content was found to be significant only during the dry period (R² = 0.48, <em>p</em> &lt; 0.05), while the effects of soil organic matter and fine roots were significant during both transition and wet periods (R² &gt; 0.5, <em>p</em> &lt; 0.05). In response to the increasing precipitation, the moderate-slope site displayed more equilibrium state of infiltration patterns than the flat and steep-slope sites, evident in the weaker correlations between the dye coverage and the lateral, preferential, and uniform flows. This result suggested that moderate-slope sites may hold greater capacity for retaining soil water. Furthermore, the linear mixed-effects models revealed significant slope and precipitation interaction influencing the infiltrating water flows (<em>p</em> &lt; 0.01). Therefore, this study provides valuable insights into the ecohydrological processes in rainforests, which are crucial for the effective planning of future forest management in tropical areas.</div></div>","PeriodicalId":12511,"journal":{"name":"Geoderma","volume":"455 ","pages":"Article 117197"},"PeriodicalIF":5.6,"publicationDate":"2025-02-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143430022","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"农林科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Digital mapping of peat thickness and extent in Finland using remote sensing and machine learning","authors":"Jonne Pohjankukka ,&nbsp;Timo A. Räsänen ,&nbsp;Timo P. Pitkänen ,&nbsp;Arttu Kivimäki ,&nbsp;Ville Mäkinen ,&nbsp;Tapio Väänänen ,&nbsp;Jouni Lerssi ,&nbsp;Aura Salmivaara ,&nbsp;Maarit Middleton","doi":"10.1016/j.geoderma.2025.117216","DOIUrl":"10.1016/j.geoderma.2025.117216","url":null,"abstract":"<div><div>Accurate data on peat extent and thickness is essential for managing drained peatlands and reducing greenhouse gas emissions. Machine learning-based digital soil mapping offers an effective approach for large-scale peat occurrence prediction. In this study, we present a workflow for producing peat occurrence maps for the whole of Finland. For this, we used random forest classification to map areas with peat thicknesses of ≥ 10 cm, ≥30 cm, ≥40 cm, and &gt; 60 cm. The input data consisted of 3.5 million point observations and 188 feature rasters from various sources. We carefully split the reference data into training and test sets, allowing for independent and robust model validation. Feature selection included an initial screening for multicollinearity using correlation-based feature pruning, followed by final selection using a genetic algorithm. Feature importance was evaluated using permutation importance and SHAP values. The resulting models utilized 26–33 features, achieving overall accuracies and F1-scores between 86–95 % and 0.82–0.95, respectively. The most important features included soil wetness indices, terrain roughness indices, and natural gamma radiation. Additionally, we provided an approach for evaluating spatial prediction uncertainty based on the models’ internal prediction agreement. Compared to existing superficial deposit maps, our peat predictions significantly improve the spatial detail of peatlands at the national level, offering new opportunities for land use planning and emission mitigation. Our exceptionally comprehensive approach is broadly applicable, offering new insights into optimizing machine learning-based digital peatland mapping, particularly through refining feature selection to account for local conditions and enhance prediction accuracy.</div></div>","PeriodicalId":12511,"journal":{"name":"Geoderma","volume":"455 ","pages":"Article 117216"},"PeriodicalIF":5.6,"publicationDate":"2025-02-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143421482","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"农林科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}