Yang Lan

Deep learning has become a key tool for carbonate thin-section image analysis. However, the lack of large public datasets limits reproducibility and fair model comparison. To address this, we present DeepCarbonate, a cleaned and standardized benchmark dataset. Samples were collected from the Ediacaran Dengying, Cambrian Longwangmiao, and Triassic Leikoupo and Jialingjiang Formations in the Sichuan Basin, China, and the Cretaceous Mishrif Formation in the UAE. The dataset was curated by petroleum geology experts; invalid images (blurred, low brightness, or corrupted) were removed through expert voting and 2σ filtering, and all images were reorganized in the ImageNet format. DeepCarbonate contains 22 lithological categories, hierarchically organized by optical mode (PPL, XPL, R) and split into train, validation, and test subsets, ensuring standardized benchmarking and reproducible experiments. Using PyTorch with CUDA acceleration, we evaluated ResNet, VGG, DenseNet, MobileNet, and EfficientNet models under baseline, ablation, long tailed distribution, and balanced Top 9 subset experiments. Results highlight the dataset’s value as a robust benchmark for carbonate petrography research and applications.

Understanding the potential impact of climate change on species distributions is crucial for biodiversity conservation and ecosystem management. Rodents, as one of the most diverse and widespread mammalian groups, play a critical role in ecological systems but also pose significant risks to agriculture systems and public health. Here, we present GridScopeRodents, a high-resolution global dataset projecting the distribution of 10 rodent genera from 2021 to 2100 under four CMIP6-based Shared Socioeconomic Pathway–Representative Concentration Pathway (SSP–RCP) scenario combinations. Using occurrence data and environmental variable, we employ the Maximum Entropy (MaxEnt) algorithm within the species distribution modeling (SDM) framework to estimate occurrence probability at a spatial resolution of 1/12° (~10 km). The dataset encompasses four SSP–RCP scenarios (SSP126, SSP245, SSP370, SSP585) and 10 global climate models (GCMs), providing projections at 20-year intervals. GridScopeRodents serves as a valuable resource for research on biodiversity conservation, invasive species monitoring, agricultural sustainability, and disease ecology. The dataset is publicly available in GeoTIFF format and can be accessed via Figshare.

Achieving sustainable development goal (SDG) 15 (life on land) depends on local conservation efforts. However, the lack of species data and inadequate ecological corridor designs significantly hinder conservation efficiency. This study focused on Sanmenxia City in China and utilized the Integrated Valuation of Ecosystem Services and Trade‐offs model, the GeoDetector model, and the Minimum Cumulative Resistance model to identify potential habitats, their influencing factors, and ecological corridors. We then adopted a new corridor feasibility model to compare the impacts of different ecological resistance, thereby identifying key biodiversity spaces. Results indicated a decline in habitat quality since 2000, primarily influenced by climate, topography, and land use. Simulation‐derived corridors using the habitat quality method were more effective than those based on expert experience. We identified 2808.45 km² of potential habitats and 20 critical corridors for land animals. This study deepens the understanding of ecological resistance impacts on corridors and provides new insights into local actions for SDG 15.

The conflict between economic growth and the arable land demand poses a significant challenge to maintaining food security and achieving the Sustainable Development Goals. Meanwhile, substantial regional disparities in food consumption contribute to variations in land demand, further exacerbating constraints on food security. However, few studies have delved into regional differences in land demand related to food consumption. To bridge these gaps, this study estimated the arable land demand and associated pressures, considering food consumption patterns and the land footprint across 31 provincial districts in China. The findings reveal that grains remain the primary crop consumed by Chinese residents. Notably, the food consumption pattern exhibits substantial disparities among provincial districts, particularly concerning livestock products. Given China's vast population and escalating consumption of livestock, the country demonstrates heightened land demands. While China does not face a national-level food security threat, regional disparities are evident, with eight provincial districts facing potential food security risks. This study explored the challenges and pathways in maintaining food security and the visions to achieve it, emphasizing the importance of sustaining a balanced food consumption pattern, reducing food waste, improving environmentally friendly agriculture practices, formulating effective and continuous laws and regulations, and exploring potential land resource development to alleviate the pressure on arable land and ensure food security.

The accurate mapping and analysis of coastal wetlands and their dynamics are crucial for local coastal wetland protection, sustainable social development, and biodiversity preservation. However, detailed mapping and comprehensive analysis of coastal wetlands remain scarce. In this study, we utilized Landsat-TM/OLI remote sensing data and employed the linear spectral mixture analysis (LSMA) method to map changes in coastal wetlands and analyze their dynamics in the Yellow River Delta (YRD) from 1991 to 2020. Our mapping results demonstrate high accuracy and are consistent with previous studies, boasting an overall accuracy exceeding 96%. During the period of 1991–2020, the YRD estuary expanded by approximately 8744.58 ha towards the east and north. The vegetation of P. australis and S. salsa underwent transformation due to agricultural practices or degradation to bare flats. Moreover, these areas saw extensive colonization by the invasive species S. alterniflora. Over the three decades, S. alterniflora expanded approximately 5 km along the coast, significantly impacting the local coastal wetland biodiversity. Furthermore, a considerable number of natural wetlands transitioned into human-made wetlands from 1991 to 2014. In particular, bare flats underwent substantial changes, transforming into aquaculture sites and salt exploitation areas. These dynamics in coastal wetlands had significant repercussions on local ecosystems, including wetland fragmentation, biodiversity depletion, and water pollution. However, post-2014, numerous wetland protection strategies were implemented, resulting in the restoration of natural wetlands. Detailed wetland mapping and dynamic analysis furnish valuable insights for the management, protection, and sustainable utilization of diverse coastal wetlands.

In response to significant shifts in dietary and lifestyle preferences, the global demand for fruits has increased dramatically, especially for apples, which are consumed worldwide. Growing apple orchards of more productive and higher quality with limited land resources is the way forward. Precise planting age identification and yield prediction are indispensable for the apple market in terms of sustainable supply, price regulation, and planting management. The planting age of apple trees significantly determines productivity, quality, and yield. Therefore, we integrated the time-series spectral endmember and logistic growth model (LGM) to accurately identify the planting age of apple orchard, and we conducted planting age-driven yield prediction using a neural network model. Firstly, we fitted the time-series spectral endmember of green photosynthetic vegetation (GV) with the LGM. By using the four-points method, the environmental carrying capacity (ECC) in the LGM was available, which serves as a crucial parameter to determine the planting age. Secondly, we combined annual planting age with historical apple yield to train the back propagation (BP) neural network model and obtained the predicted apple yields for 12 counties. The results show that the LGM method can accurately estimate the orchard planting age, with Mean Absolute Error (MAE) being 1.76 and the Root Mean Square Error (RMSE) being 2.24. The strong correlation between orchard planting age and apple yield was proved. The results of planting age-driven yield prediction have high accuracy, with the MAE up to 2.95% and the RMSE up to 3.71%. This study provides a novel method to accurately estimate apple orchard planting age and yields, which can support policy formulation and orchard planning in the future.

The carbonate reservoirs in the Middle Permian Maokou Formation have been proved to be outstanding gas exploration targets. Previous researches unravel carbonate reservoirs are mainly developed in Maokou-2 and Maokou-3 Members, where Maokou-1 Member is mostly source rock. However, with the discovery of unconventional gas reservoirs in Maokou-1 Member, evaluation of porosity, permeability, TOC, and lithology in Maokou-1 Member in Sichuan basin is becoming increasingly essential. This study designs a new multi task XGBoost (MT-XGBoost) model with hard parameter sharing framework and compares it with traditional Random Forest (RF), XGBoost (XGBoost) and Artificial Neural Network models (ANN) in single tasks (all models reaches SOTA). Result shows the accuracies of MT-XGBoost for lithology, porosity, permeability, and TOC are 91.5%, 96%, 97% and 93%, respectively. For XGBoost, the accuracies are 81.25%, 92%, 97% and 93%78.2%. For RF, the accuracies are 68.5%, 95%, 96% and 91%. For ANN, the accuracies are 68%, 76%, 75% and 80%. To test MT-XGBoost, this study collects data from YF66-1 and ZG1 (outside the research area) and puts them into MT-XGBoost. The results in YF66-1 and ZG1 proves the generalization capability of MT-XGBoost. Predictions indicate central and southeastern areas are the most potential zones for unconventional gas exploration in the northeastern Sichuan basin. MT-XGBoost could be utilized in the future evaluation of Maokou-1 Member in the whole Sichuan basin.