Revisiting the 2023 wildfire season in Canada

Science of Remote Sensing, Год журнала: 2024, Номер 10, С. 100145 - 100145

Опубликована: Июнь 21, 2024

The area burned by wildfires in Canada 2023 is unprecedented historical records. To help ensure the safety of communities and support mobilization firefighting resources, rapid detection areas affected required. Satellite data are ideally suited to provide near real-time wildfire information over large areas. At same time, clouds, smoke, haze can obscure collection observations from sensors typically used for mapping purposes. Established methods using coarse spatial resolution satellites (e.g., MODIS, VIIRS) rely upon combination daily revisit enable reliable active fires, full or part, application modeling (including buffering) infer additional, yet still obscured, While timely, these initial maps wildfire-impacted do not capture small fires (those smaller than 200 ha) and, importantly, intended differentiate unburned within fire perimeters. address limitations, we Sentinel-2A -2B, Landsat-8 -9, which form a virtual constellation four map Canada's forested ecosystems season. Availing high temporal density Tracking Intra- Inter-year Change algorithm (TIIC), an aggregate seasonal resulted total 12.74 Mha. Within this area, 9.51 Mha treed land cover was impacted. Shrubs wetlands comprised most remaining non-treed that burned. Using 2022 aboveground biomass (AGB), approximately 0.649 Pg AGB impacted wildfires, representing 11-fold increase impacts relative long-term annual average loss. Differences between estimate reported herein indicated Natural Resources (NRCan) Fire M3 hotspot perimeters (18.64 Mha) were analyzed. Overall, estimates differed 5.9 Mha, including 1.13 water identified as NRCan two products also investigated quantified. TIIC enables near-continuous through season, allowing within-year refinement interrogation types impacted, estimation associated consequences.

Язык: Английский

---

Advancing horizons in remote sensing: a comprehensive survey of deep learning models and applications in image classification and beyond

Neural Computing and Applications, Год журнала: 2024, Номер 36(27), С. 16727 - 16767

Опубликована: Авг. 2, 2024

Abstract In recent years, deep learning has significantly reshaped numerous fields and applications, fundamentally altering how we tackle a variety of challenges. Areas such as natural language processing (NLP), computer vision, healthcare, network security, wide-area surveillance, precision agriculture have leveraged the merits era. Particularly, improved analysis remote sensing images, with continuous increase in number researchers contributions to field. The high impact development is complemented by rapid advancements availability data from sensors, including high-resolution RGB, thermal, LiDAR, multi-/hyperspectral cameras, well emerging platforms satellites aerial vehicles that can be captured multi-temporal, multi-sensor, devices wider view. This study aims present an extensive survey encapsulates widely used strategies for tackling image classification challenges sensing. It encompasses exploration imaging platforms, sensor varieties, practical prospective developments

Язык: Английский

---

Generation of 100-m, Hourly Land Surface Temperature Based on Spatio-Temporal Fusion

IEEE Transactions on Geoscience and Remote Sensing, Год журнала: 2024, Номер 62, С. 1 - 16

Опубликована: Янв. 1, 2024

Landsat surface temperature (LST) is an important physical quantity for global climate change monitoring. Over the past decades, several LST products have been produced by satellite thermal infrared (TIR) bands or land models (LSMs). Recent research has increased spatio-temporal resolution of to 2 km, hourly based on Geostationary Operational Environmental Satellites (GOES)-R Advanced Baseline Imager (ABI) data. The spatial however, insufficient monitoring at regional scale. This paper investigates feasibility applying fusion generate reliable 100 m, data newly released GOES-16 ABI and m most accurate method was identified through a comparison between popular methods. Furthermore, comprehensive performed (with LST) involving satellite-derived (i.e., GOES) model-derived LSMs European Centre Medium-range Weather Forecasts (ECMWF) Reanalysis v .5 (ERA5)-Land). temporal adaptive reflectance model (STARFM) demonstrated be appropriate data, which average root mean square error (RMSE) 2.640 K, absolute (MAE) 2.159 K coefficient determination ( xmlns:xlink="http://www.w3.org/1999/xlink">R ² ) 0.982 referring xmlns:xlink="http://www.w3.org/1999/xlink">in situ time-series. inheriting advantages direct observation, GOES generation greater accuracy compared ERA5-Land in experiments. generated can provide diurnal with fine various applications.

Язык: Английский

---

China’s ongoing rural to urban transformation benefits the population but is not evenly spread

Communications Earth & Environment, Год журнала: 2024, Номер 5(1)

Опубликована: Авг. 4, 2024

China prioritizes a coordinated and sustainable shift from rural to urban areas, termed rural-urban transformation. This involves land, population, industry urbanization. Here we explore the spatiotemporal dynamics of transformation patterns in China, focusing on degree integrated coupling between three tracks. To conduct our investigation, utilized urbanization cube theory, satellite-derived gridded datasets, self-organizing map. Our findings show that eastern has higher levels compared western China. There been an overall increase China's We identified six typical across Over time, 53.58% prefectures improved patterns, 3.44% degraded, 42.98% (mainly China) remained unchanged. More importantly, highlight increasing reduced inequities well-being. The rural-to-urban integrates changes land use, development reduces well-being is more evident East but not West according analysis combines satellite data, statistical analysis, machine learning.

Язык: Английский

---

Satellite Remote Sensing of Savannas: Current Status and Emerging Opportunities

Journal of Remote Sensing, Год журнала: 2022, Номер 2022

Опубликована: Янв. 1, 2022

Savannas cover a wide climatic gradient across large portions of the Earth’s land surface and are an important component terrestrial biosphere. have been undergoing changes that alter composition structure their vegetation such as encroachment woody increasing land-use intensity. Monitoring spatial temporal dynamics savanna ecosystem (e.g., partitioning herbaceous vegetation) function aboveground biomass) is high importance. Major challenges include misclassification savannas forests at mesic end range, disentangling contribution to biomass, quantifying mapping fuel loads. Here, we review current (2010–present) research in application satellite remote sensing regional global scales. We identify emerging opportunities can help overcome existing challenges. provide recommendations on how these be leveraged, specifically (1) development conceptual framework leads consistent definition sensing; (2) improving ecologically relevant information soil properties fire activity; (3) exploiting high-resolution imagery provided by nanosatellites better understand role landscape functioning; (4) using novel approaches from artificial intelligence machine learning combination with multisource observations, e.g., multi-/hyperspectral, synthetic aperture radar (SAR), light detection ranging (lidar), data plant traits infer potentially new relationships between biotic abiotic components either proven or disproven targeted field experiments.

Язык: Английский

---