Global trends and regime state shifts of lacustrine aquatic vegetation

The Innovation, Год журнала: 2025, Номер 6(3), С. 100784 - 100784

Опубликована: Янв. 18, 2025

Aquatic vegetation (AV) is vital for maintaining the health of lake ecosystems, with submerged aquatic (SAV) and floating/emergent (FEAV) representing clear shaded states, respectively. However, global SAV FEAV dynamics are poorly understood due to data scarcity. To address this gap, we developed an innovative AV mapping algorithm workflow using satellite imagery (1.4 million Landsat images) from 1989 2021 created a database across 5,587 shallow lakes. Our findings suggest that covers 108,186 km2 on average globally, accounting 28.9% (FEAV, 15.8%; SAV, 13.1%) total area. Over two decades, observed notable transition: decreased by 30.4%, while increased 15.6%, leading substantial net loss AV. This trend indicates shift conditions, increasingly progressing toward turbid states dominated phytoplankton. We found human-induced eutrophication was primary driver change until early 2010s, after which warming rising temperatures became dominant drivers. These trends serve as warning sign deteriorating worldwide. With future climate intensified eutrophication, these ongoing pose significant risk disrupting ecosystems.

Язык: Английский

---

Aboveground biomass retrieval of wetland vegetation at the species level using UAV hyperspectral imagery and machine learning

Ecological Indicators, Год журнала: 2024, Номер 166, С. 112365 - 112365

Опубликована: Июль 13, 2024

Язык: Английский

---

Supervised and unsupervised machine learning approaches for tree classification using multiwavelength airborne polarimetric LiDAR

Smart Agricultural Technology, Год журнала: 2025, Номер unknown, С. 100872 - 100872

Опубликована: Март 1, 2025

Язык: Английский

---

Detailed wetland-type classification using Landsat-8 time-series images: a pixel- and object-based algorithm with knowledge (POK)

GIScience & Remote Sensing, Год журнала: 2023, Номер 61(1)

Опубликована: Дек. 15, 2023

Mapping detailed wetland types can offer useful information for management and protection, which strongly support the Global Biodiversity Framework. Many studies have conducted classification at regional, national, global scale, whereas fine-resolution mapping with is still challenging. To address this issue, we developed an integration of pixel- object-based algorithms knowledge (POK) by combining pixel-based random forest hierarchical decision tree. Taking Guangxi Beibu Gulf Economic Zone (GBGEZ) Guangdong-Hong Kong-Macao Greater Bay Area (GBA) as our study areas, produced maps 10 6 non-wetland using Landsat-8 time series. In addition, to comprehensively evaluate accuracy classification, implemented validation based on test samples data inter-comparison existing datasets, respectively. The results indicate that overall map was 91.6% ±1.2%. For types, agricultural pond, coastal shallow water, floodplain, mangrove, reservoir, river, tidal flat achieved good accuracies, both user producer exceeding 88.0%. most accuracies were greater than 72.0%. By comparison it found had consistencies China Ecosystem-type Classification Dataset (CECD) land use dataset, MC_LASAC mangrove Tidal Wetlands in East Asia (TWEA) dataset. 2020, area 4,198.8 km2 GBGEZ 10,932.2 GBA. main two urban agglomerations ponds, waters, mangroves, reservoirs, rivers, flats. Our successfully mapped GBA, serving Framework Convention Biological Diversity.

Язык: Английский

---

Evolution of Small and Micro Wetlands and Their Driving Factors in the Yangtze River Delta—A Case Study of Wuxi Area

Remote Sensing, Год журнала: 2023, Номер 15(4), С. 1152 - 1152

Опубликована: Фев. 20, 2023

Understanding the long-term dynamics and driving factors behind small micro wetlands is critical for their management future sustainability. This study explored impacts of natural anthropogenic on spatiotemporal evolution these areas in Wuxi area using support vector machine (SVM) classification method geographic detector model based Landsat satellite image data from 1985 to 2020. The results revealed that: (1) Natural were prominent area, with an average proportion 70%, although they exhibited a downward trend over last ten years, scale increased 1.5-fold—from 4349.59 hm2 10,841.59 (2) had obvious seasonal variations, most being 0.1–1 1–3 hm2, respectively. From perspective spatial distribution, primarily distributed Yixing district, which accounts 34% area. (3) distribution was systematically affected by human activities. main that annual temperature GDP, interactions between all nonlinear bi-linear. influences weakened, activities steadily emerging as dominant factor distribution. this can provide supportive scientific basis ecological restoration protection wetlands.

Язык: Английский

---

Automated Tree Detection Using Image Processing and Multisource Data

Applied Sciences, Год журнала: 2025, Номер 15(2), С. 667 - 667

Опубликована: Янв. 11, 2025

This paper presents a method for the automatic detection and assessment of trees tree-covered areas in Katowice, capital Upper Silesian Industrial Region southern Poland. The proposed approach utilizes satellite imagery height maps, employing image-processing techniques integrating data from various sources. We developed pipeline gathering pre-processing information, including vegetation numerical land-cover models, which were used to derive new tree detection. Our findings confirm that can significantly enhance efficiency urban management processes, contributing creation greener more resident-friendly cities.

Язык: Английский

---

A Modified Swin-UNet Model for Coastal Wetland Classification Using Multi-Temporal Sentinel-2 Images

Estuaries and Coasts, Год журнала: 2025, Номер 48(3)

Опубликована: Фев. 24, 2025

Язык: Английский

---

Deep-Learning-Based Automatic Extraction of Aquatic Vegetation from Sentinel-2 Images—A Case Study of Lake Honghu

Remote Sensing, Год журнала: 2024, Номер 16(5), С. 867 - 867

Опубликована: Фев. 29, 2024

Aquatic vegetation is an important component of aquatic ecosystems; therefore, the classification and mapping aspect lake management. Currently, decision tree (DT) method based on spectral indices has been widely used in extraction data, but disadvantage this that it difficult to fix threshold value, which, turn, affects automatic effect. In study, Sentinel-2 MSI data were produce a sample set (about 930 samples) four inland lakes (Lake Taihu, Lake Caohai, Honghu, Dongtinghu) using visual interpretation method, including emergent, floating-leaved, submerged vegetation. Based set, DL model (Res-U-Net) was train model. The achieved higher overall accuracy, relevant error, kappa coefficient (90%, 8.18%, 0.86, respectively) compared DT (79%, 23.07%, 0.77) random forest (78%,10.62% when utilizing results as ground truth. When measured point truth, exhibited accuracies 59%, 78%, 91% for submerged, emergent vegetation, respectively. addition, still maintains good recognition presence clouds with influence water bloom. applying Honghu from January 2017 October 2023, obtained temporal variation patterns consistent other studies. study paper shows proposed application potential extracting data.

Язык: Английский

---

Estimation aboveground biomass in subtropical bamboo forests based on an interpretable machine learning framework

Environmental Modelling & Software, Год журнала: 2024, Номер 178, С. 106071 - 106071

Опубликована: Май 10, 2024

Forest biomass is an essential indicator of forest ecosystem carbon cycle and global climate change research, traditional machine learning cannot explain the mechanism feature variable impact on aboveground (AGB). Therefore, we proposed interpretable bamboo AGB prediction method based Shaply Additive exPlanation (SHAP) XGBoost model to variables AGB. The estimated using monthly annual scale leaf area index (LAI), enhanced vegetation (EVI), ratio (RVI), precipitation (Pre), maximum temperature (Tmax), minimum (Tmin) solar radiation (Rad) data. results showed that could be effectively predict AGB, more important than temperature. framework revealed threshold effect, exceeded value, impacts LAI_Ann, EVI_Ann, Pre_11 were stable. SHAP interaction value between LAI_Ann EVI_Ann decreased with increasing LAI_Ann. By contrast, when increased, increased also easily implemented, providing

Язык: Английский

---

Mapping global lake aquatic vegetation dynamics using 10-m resolution satellite observations

Science Bulletin, Год журнала: 2024, Номер 69(19), С. 3115 - 3126

Опубликована: Май 17, 2024

Язык: Английский

---