Urban agriculture offers an alternative land use for integrating multiple functions in densely populated areas. While urban agriculture has historically been an important element of cities in many developing countries, recent concerns about economic and food security have resulted in a growing movement to produce food in cities of developed countries including the United States. In these regions, urban agriculture offers a new frontier for land use planners and landscape designers to become involved in the development and transformation of cities to support community farms, allotment gardens, rooftop gardening, edible landscaping, urban forests, and other productive features of the urban environment. Despite the growing interest in urban agriculture, urban planners and landscape designers are often ill-equipped to integrate food-systems thinking into future plans for cities. The challenge (and opportunity) is to design urban agriculture spaces to be multifunctional, matching the specific needs and preferences of local residents, while also protecting the environment. This paper provides a review of the literature on urban agriculture as it applies to land use planning in the United States. The background includes a brief historical perspective of urban agriculture around the world, as well as more recent examples in the United States. Land use applications are considered for multiple scales, from efforts that consider an entire city, to those that impact a single building or garden. Barriers and constraints to urban agriculture are discussed, followed by research opportunities and methodological approaches that might be used to address them. This work has implications for urban planners, landscape designers, and extension agents, as opportunities to integrate urban agriculture into the fabric of our cities expand.
Plantations are established for a variety of reasons including wood production, soil and water conservation, and more recently, carbon sequestration. The effect of this growing land-use change on biodiversity, however, is poorly understood and considerable debate exists as to whether plantations are ‘green deserts’ or valuable habitat for indigenous flora and fauna. This paper synthesizes peer-reviewed articles that provide quantitative data on plant species richness in plantations and paired land uses, most often representative of pre-plantation land cover. The results of this synthesis suggest that the value of plantations for biodiversity varies considerably depending on whether the original land cover is grassland, shrubland, primary forest, secondary forest, or degraded or exotic pasture, and whether native or exotic tree species are planted. The results of this study suggest that plantations are most likely to contribute to biodiversity when established on degraded lands rather than replacing natural ecosystems, such as forests, grasslands, and shrublands, and when indigenous tree species are used rather than exotic species. These findings can help guide afforestation and reforestation programs, including those aimed at increasing terrestrial carbon sequestration.
Global climate change and intensified human activities drive rapid land use and land cover (LULC) changes, particularly in ecologically fragile regions like China's Southern Hilly Region (SHR), affecting ecosystem services (ESs) trade-offs/synergies. However, scale-dependent thresholds governing these relationships remain poorly quantified. We analyzed ESs dynamics (water yield, soil conservation, carbon storage, nutrient retention, habitat quality) across regional, watershed, and sub-watershed scales (1990–2020) using the InVEST model, sensitivity indices, and piecewise linear regression. ESs responses exhibited significant scale effects, with sub-watersheds showing the highest sensitivity to LULC changes and representing the most stable management unit. Particularly, we found that critical LULC thresholds regulate trade-offs/synergies: forest cover exceeding ∼70 % strongly enhanced synergies among multiple ESs, while cropland proportions between 30 and 65 % intensified trade-offs (e.g., between soil conservation and water yield). Impervious expansion consistently degraded ESs. Our results demonstrate that optimizing LULC patterns-prioritizing forest conservation (>70 % cover), limiting cropland (<65 %), and controlling urban sprawl-at the sub-watershed scale minimizes ESs trade-offs. This study establishes quantitative thresholds to guide targeted land-use planning and ecological restoration policies in hilly regions globally, supporting sustainable landscape governance.
Abstract Background This study used an online survey to explore the perspectives, practices and knowledge gaps of Irish farmers regarding the adoption of multispecies swards (MSS), a sustainable alternative to traditional monoculture grassland systems. With ruminant livestock production being central to global agricultural gross domestic product and Ireland's reliance on grass‐based systems, MSS offer potential benefits for productivity, sustainability and environmental impact. However, farm‐level data on MSS adoption are limited. Methods An adapted version of Rogers' Innovation Decision Process model was used to examine farmers' awareness, adoption drivers, perceived benefits, barriers and knowledge needs related to MSS. Results Among 200 Irish farmers surveyed between October 2023 and March 2024, 93% were aware of MSS and 57% had adopted it. Reported benefits included improved biodiversity, soil health, drought resilience and reduced nitrogen use, with 91% of adopters lowering fertiliser inputs. Key barriers were difficulties with establishment, grazing management, weed control and uncertainty about seed mixtures. Farmers expressed a need for more guidance on persistence and management and preferred learning via open days and discussion groups. Conclusions The findings highlight the need for tailored support to facilitate MSS adoption. Future initiatives should prioritise peer learning, demonstration farms and practical guidance on establishment and grazing.
Shaoming Han, Cheng Qian, Nawal Abdalla Adam
et al.
This study examines the impact of financial inclusion on bank stability across 36 emerging economies, utilizing bank-level data from over 1,500 commercial banks spanning the period 2004 to 2023. Despite the recognized benefits of financial inclusion, its influence on banking stability remains complex and context dependent. The research employs advanced econometric methodologies, including fixed-effects models, Driscoll-Kraay standard errors to address heteroskedasticity and cross-sectional dependence, and system Generalized Method of Moments (GMM) estimation to control for endogeneity and dynamic effects. The findings reveal that financial inclusion generally enhances bank stability and positively influences operational efficiency and funding stability. However, during periods of lax financial regulations or excessive government intervention, banks may engage in riskier behaviors, potentially undermining stability. Key results indicate that (1) robust economic growth and stable policy environments amplify the positive effects of financial inclusion on bank stability, (2) excessive government control may foster risk-taking behaviors, (3) strong financial conditions mitigate adverse impacts, (4) financial inclusion improves risk management and operational efficiency, and (5) effective regulatory frameworks are pivotal in leveraging financial inclusion for sound banking operations. These insights suggest that policymakers in emerging markets should carefully balance the promotion of financial inclusion with safeguards that maintain financial stability.
Semantic segmentation of ultra-high-resolution remote sensing (UHR-RS) imagery plays a critical role in land use and land cover analysis, yet it remains computationally intensive due to the enormous input size and high spatial complexity. Existing studies have commonly employed strategies such as patch-wise processing, multi-scale model architectures, lightweight networks, and representation sparsification to reduce resource demands, but they have often struggled to maintain long-range contextual awareness and scalability for inputs of arbitrary size. To address this, we propose RingFormer-Seg, a scalable Vision Transformer framework that enables long-range context learning through multi-device parallelism in UHR-RS image segmentation. RingFormer-Seg decomposes the input into spatial subregions and processes them through a distributed three-stage pipeline. First, the Saliency-Aware Token Filter (STF) selects informative tokens to reduce redundancy. Next, the Efficient Local Context Module (ELCM) enhances intra-region features via memory-efficient attention. Finally, the Cross-Device Context Router (CDCR) exchanges token-level information across devices to capture global dependencies. Fine-grained detail is preserved through the residual integration of unselected tokens, and a hierarchical decoder generates high-resolution segmentation outputs. We conducted extensive experiments on three benchmarks covering UHR-RS images from 2048 × 2048 to 8192 × 8192 pixels. Results show that our framework achieves top segmentation accuracy while significantly improving computational efficiency across the DeepGlobe, Wuhan, and Guangdong datasets. RingFormer-Seg offers a versatile solution for UHR-RS image segmentation and demonstrates potential for practical deployment in nationwide land cover mapping, supporting informed decision-making in land resource management, environmental policy planning, and sustainable development.
Land surface temperature (LST) is vital for land-atmosphere interactions and climate processes. Accurate LST retrieval remains challenging under heterogeneous land cover and extreme atmospheric conditions. Traditional split window (SW) algorithms show biases in humid environments; purely machine learning (ML) methods lack interpretability and generalize poorly with limited data. We propose a coupled mechanism model-ML (MM-ML) framework integrating physical constraints with data-driven learning for robust LST retrieval. Our approach fuses radiative transfer modeling with data components, uses MODTRAN simulations with global atmospheric profiles, and employs physics-constrained optimization. Validation against 4,450 observations from 29 global sites shows MM-ML achieves MAE=1.84K, RMSE=2.55K, and R-squared=0.966, outperforming conventional methods. Under extreme conditions, MM-ML reduces errors by over 50%. Sensitivity analysis indicates LST estimates are most sensitive to sensor radiance, then water vapor, and less to emissivity, with MM-ML showing superior stability. These results demonstrate the effectiveness of our coupled modeling strategy for retrieving geophysical parameters. The MM-ML framework combines physical interpretability with nonlinear modeling capacity, enabling reliable LST retrieval in complex environments and supporting climate monitoring and ecosystem studies.
Geomorphometric modeling and mapping of ice-free Antarctic areas can be applied for obtaining new quantitative knowledge about the topography of these unique landscapes and for the further use of morphometric information in Antarctic research. Within the framework of a project of creating a physical geographical thematic scientific reference geomorphometric atlas of ice-free areas of Antarctica, we performed geomorphometric modeling and mapping of five key coastal oases of Enderby Land, East Antarctica. These include, from west to east, the Konovalov Oasis, Thala Hills (Molodezhny and Vecherny Oases), Fyfe Hills, and Howard Hills. As input data, we used five fragments of the Reference Elevation Model of Antarctica (REMA). For the coastal oases and adjacent ice sheet and glaciers, we derived models and maps of eleven, most scientifically important morphometric variables (i.e., slope, aspect, horizontal curvature, vertical curvature, minimal curvature, maximal curvature, catchment area, topographic wetness index, stream power index, total insolation, and wind exposition index). In total, we derived 60 maps in 1:50,000 and 1:75,000 scales. The obtained models and maps describe the coastal oases of Enderby Land in a rigorous, quantitative, and reproducible manner. New morphometric data can be useful for further geological, geomorphological, glaciological, ecological, and hydrological studies of this region.
Saumya Malik, Valentina Pyatkin, Sander Land
et al.
Reward models are used throughout the post-training of language models to capture nuanced signals from preference data and provide a training target for optimization across instruction following, reasoning, safety, and more domains. The community has begun establishing best practices for evaluating reward models, from the development of benchmarks that test capabilities in specific skill areas to others that test agreement with human preferences. At the same time, progress in evaluation has not been mirrored by the effectiveness of reward models in downstream tasks -- simpler direct alignment algorithms are reported to work better in many cases. This paper introduces RewardBench 2, a new multi-skill reward modeling benchmark designed to bring new, challenging data for accuracy-based reward model evaluation -- models score about 20 points on average lower on RewardBench 2 compared to the first RewardBench -- while being highly correlated with downstream performance. Compared to most other benchmarks, RewardBench 2 sources new human prompts instead of existing prompts from downstream evaluations, facilitating more rigorous evaluation practices. In this paper, we describe our benchmark construction process and report how existing models perform on it, while quantifying how performance on the benchmark correlates with downstream use of the models in both inference-time scaling algorithms, like best-of-N sampling, and RLHF training algorithms like proximal policy optimization.