A spatio-temporal analysis investigating completeness and inequalities of global urban building data in OpenStreetMap

Nature Communications, Journal Year: 2023, Volume and Issue: 14(1)

Published: July 6, 2023

Abstract OpenStreetMap (OSM) has evolved as a popular dataset for global urban analyses, such assessing progress towards the Sustainable Development Goals. However, many analyses do not account uneven spatial coverage of existing data. We employ machine-learning model to infer completeness OSM building stock data 13,189 agglomerations worldwide. For 1,848 centres (16% population), footprint exceeds 80% completeness, but remains lower than 20% 9,163 cities (48% population). Although inequalities have recently receded, partially result humanitarian mapping efforts, complex unequal pattern biases remains, which vary across various human development index groups, population sizes and geographic regions. Based on these results, we provide recommendations producers analysts manage data, well framework support assessment biases.

Language: Английский

---

Mapping the presence and distribution of tree species in Canada's forested ecosystems

Remote Sensing of Environment, Journal Year: 2022, Volume and Issue: 282, P. 113276 - 113276

Published: Sept. 29, 2022

Knowledge of tree species is required to inform management, planning, and monitoring forests as well characterize habitat ecosystem function. Remotely sensed data spatial modeling enable mapping presence distribution. Following an assessment identified in the sample-based National Forest Inventory (NFI), we mapped 37 over 650-Mha, forest-dominated ecosystems Canada representing 2019 conditions. Landsat imagery related spectral indices, geographic climate data, elevation derivatives, remote sensing-derived phenology are used predictor variables trained with calibration samples from Canada's NFI using Random Forests machine learning algorithm. Based upon prior knowledge distributions, classification models were implemented on a regional basis, meaning only that expected given region modeled local samples. Modeling resulted class membership probabilities values for each regionally eligible all treed pixels indicator attribution confidence derived distance feature space between two leading classes. Accuracy was conducted independent validation also drawn following same selection rules indicated overall accuracy 93.1% ± 0.1% (95%-confidence interval). Predictor informing geographic, climatic topographic conditions had largest importance models. Nationally, most common black spruce (Picea mariana; 203 Mha or 57.3% area), trembling aspen (Populus tremuloides; 34.7 Mha, 9.8%), lodgepole pine (Pinus contorta; 21.1 5.9%). Regionally, there ecozone-level dominance other species, including subalpine fir (Abies lasiocarpa; Montane Cordillera), western hemlock (Tsuga heterophylla; Pacific Maritime), balsam balsamea; Atlantic Maritime). per-pixel probabilities, assemblages akin those forest inventories can be produced. Further, calibrated reflectance imagery, methods presented herein time series images. The approach uses open variables, making method portable areas availability training data.

Language: Английский

---

The neglected role of abandoned cropland in supporting both food security and climate change mitigation

Nature Communications, Journal Year: 2023, Volume and Issue: 14(1)

Published: Sept. 28, 2023

Despite the looming land scarcity for agriculture, cropland abandonment is widespread globally. Abandoned can be reused to support food security and climate change mitigation. Here, we investigate potentials trade-offs of using global abandoned recultivation restoring forests by natural regrowth, with spatially-explicit modelling scenario analysis. We identify 101 Mha between 1992 2020, a capability concurrently delivering 29 363 Peta-calories yr

Language: Английский

---

Assessing and improving the transferability of current global spatial prediction models

Global Ecology and Biogeography, Journal Year: 2023, Volume and Issue: 32(3), P. 356 - 368

Published: Jan. 26, 2023

Abstract Aim Global‐scale maps of the environment are an important source information for researchers and decision makers. Often, these created by training machine learning algorithms on field‐sampled reference data using remote sensing as predictors. Since field samples often sparse clustered in geographic space, model prediction requires a transfer trained to regions where no available. However, recent studies question feasibility predictions far beyond location data. Innovation We propose novel workflow spatial predictive mapping that leverages developments this combines them innovative ways with aim improved transferability performance assessment. demonstrate, evaluate discuss from recently published global environmental maps. Main conclusions Reducing predictors those relevant leads increase map accuracy without decrease quality areas high sampling density. Still, reliable gap‐free were not possible, highlighting their evaluation hampered limited availability

Language: Английский

---

Emergent temperature sensitivity of soil organic carbon driven by mineral associations

Nature Geoscience, Journal Year: 2024, Volume and Issue: 17(3), P. 205 - 212

Published: Feb. 20, 2024

Abstract Soil organic matter decomposition and its interactions with climate depend on whether the is associated soil minerals. However, data limitations have hindered global-scale analyses of mineral-associated particulate carbon pools their benchmarking in Earth system models used to estimate cycle–climate feedbacks. Here we analyse observationally derived global estimates quantify relative proportions compute climatological temperature sensitivities as decline increasing temperature. We find that sensitivity average 28% higher than carbon, up 53% cool climates. Moreover, distribution between these underlying drives emergent bulk stocks. vary widely predictions pool distributions. show proportion model are conceptually similar mineral-protected ranges from 16 85% across Coupled Model Intercomparison Project Phase 6 offline land models, implications for ages ecosystem responsiveness. To improve projections feedbacks, it imperative assess accurately predict vulnerability carbon.

Language: Английский

---

Benthic habitat mapping: A review of three decades of mapping biological patterns on the seafloor

Estuarine Coastal and Shelf Science, Journal Year: 2023, Volume and Issue: 296, P. 108599 - 108599

Published: Dec. 12, 2023

What is benthic habitat mapping, how it accomplished, and has that changed over time? We query the published literature to answer these questions synthesize results quantitatively provide a comprehensive review of field past three decades. Categories maps are differentiated unambiguously by response variable (i.e., subject being mapped) rather than approaches used produce map. Additional terminology in clarified defined based on provenance, statistical criteria, common usage. Mapping approaches, models, data sets, technologies, range other attributes reviewed their application, we document changes ways components have been integrated map habitats time. found use acoustic remote sensing surpassed optical methods for obtaining environmental data. Although wide variety employed ground truth maps, underwater imagery become most validation tool – surpassing physical sampling. The empirical machine learning models process increased dramatically 10 years, superseded expert manual interpretation. discuss products derived from address ecological emerging seascape ecology, technologies survey logistics pose different challenges this research across ecosystems intertidal shallow sublittoral regions deep ocean. Outstanding identified discussed context with trajectory field.

Language: Английский

---

Uncertainty-aware deep learning for reliable health monitoring in safety-critical energy systems

Energy, Journal Year: 2024, Volume and Issue: 291, P. 130419 - 130419

Published: Jan. 21, 2024

Language: Английский

---

How Interpretable Machine Learning Can Benefit Process Understanding in the Geosciences

Earth s Future, Journal Year: 2024, Volume and Issue: 12(7)

Published: July 1, 2024

Abstract Interpretable Machine Learning (IML) has rapidly advanced in recent years, offering new opportunities to improve our understanding of the complex Earth system. IML goes beyond conventional machine learning by not only making predictions but also seeking elucidate reasoning behind those predictions. The combination predictive power and enhanced transparency makes a promising approach for uncovering relationships data that may be overlooked traditional analysis. Despite its potential, broader implications field have yet fully appreciated. Meanwhile, rapid proliferation IML, still early stages, been accompanied instances careless application. In response these challenges, this paper focuses on how can effectively appropriately aid geoscientists advancing process understanding—areas are often underexplored more technical discussions IML. Specifically, we identify pragmatic application scenarios typical geoscientific studies, such as quantifying specific contexts, generating hypotheses about potential mechanisms, evaluating process‐based models. Moreover, present general practical workflow using address research questions. particular, several critical common pitfalls use lead misleading conclusions, propose corresponding good practices. Our goal is facilitate broader, careful thoughtful integration into science research, positioning it valuable tool capable enhancing current

Language: Английский

---

Explainable artificial intelligence and interpretable machine learning for agricultural data analysis

Artificial Intelligence in Agriculture, Journal Year: 2022, Volume and Issue: 6, P. 257 - 265

Published: Jan. 1, 2022

Artificial intelligence and machine learning have been increasingly applied for prediction in agricultural science. However, many models are typically black boxes, meaning we cannot explain what the learned from data reasons behind predictions. To address this issue, I introduce an emerging subdomain of artificial intelligence, explainable (XAI), associated toolkits, interpretable learning. This study demonstrates usefulness several methods by applying them to openly available dataset. The dataset includes no-tillage effect on crop yield relative conventional tillage soil, climate, management variables. Data analysis discovered that can increase maize where is <5000 kg/ha maximum temperature higher than 32°. These useful answer (i) which variables important regression/classification, (ii) variable interactions prediction, (iii) how their with response variable, (iv) underlying a predicted value certain instance, (v) whether different algorithms offer same these questions. argue goodness model fit overly evaluated performance measures current practice, while questions unanswered. XAI enhance trust explainability AI.

Language: Английский

---

Vegetation structure derived from airborne laser scanning to assess species distribution and habitat suitability: The way forward

Diversity and Distributions, Journal Year: 2022, Volume and Issue: 29(1), P. 39 - 50

Published: Oct. 30, 2022

Abstract Ecosystem structure, especially vertical vegetation is one of the six essential biodiversity variable classes and an important aspect habitat heterogeneity, affecting species distributions diversity by providing shelter, foraging, nesting sites. Point clouds from airborne laser scanning (ALS) can be used to derive such detailed information on structure. However, public agencies usually only provide digital elevation models, which do not Calculating structure variables ALS point requires extensive data processing remote sensing skills that most ecologists have. extremely valuable for many analyses use distribution. We here propose 10 should easily accessible researchers stakeholders through national portals. In addition, we argue a consistent selection their systematic testing, would allow continuous improvement list keep it up‐to‐date with latest evidence. This initiative particularly needed advance ecological research open datasets but also guide potential users in face increasing availability global products.

Language: Английский

---