Wetland Vegetation Research Articles

Vegetation and carbon sink response to water level changes in a seasonal lake wetland

Water level fluctuations are among the main factors affecting the development of wetland vegetation communities, carbon sinks, and ecological processes. Hongze Lake is a typical seasonal lake wetland in the Huaihe River Basin. Its water levels have experienced substantial fluctuations because of climate change, as well as gate and dam regulations. In this study, long-term cloud-free remote sensing images of water body area, net plant productivity (NPP), gross primary productivity (GPP), and Fractional vegetation cover (FVC) of the wetlands of Hongze Lake were obtained from multiple satellites by Google Earth Engine (GEE) from 2006 to 2023. The trends in FVC were analyzed using a combined Theil-Sen estimator and Mann-Kendall (MK) test. Linear regression was employed to analyze the correlation between the area of water bodies and that of different degrees of FVC. Additionally, annual frequencies of various water levels were constructed to explore their association with GPP, NPP, and FVC.The results showed that water level fluctuations significantly influence the spatial and temporal patterns of wetland vegetation cover and carbon sinks, with a significant correlation (P<0.05) between water levels and vegetation distribution. Following extensive restoration efforts, the carbon sink capacity of the Hongze Lake wetland has increased. However, it is essential to consider the carbon sink capacity in areas with low vegetation cover, for the lakeshore zone with a higher inundation frequency and low vegetation cover had a lower carbon sink capacity. These findings provide a scientific basis for the establishment of carbon sink enhancement initiatives, restoration programs, and policies to improve the ecological value of wetland ecosystem conservation areas.

Impact of land use/land cover (LU/LC) changes on ecosystem service values in Muger Sub-basin, Upper Blue Nile Basin, Ethiopia

This study was conducted to estimate the impact of land use/land cover (LULC) changes on the ecosystem service value (ESV) in the Upper Blue Nile basin over three-decade (1986–2020). The findings revealed that there is a decline in total ESV from US$ 492.77 × 106 in 1986 to US$ 400.65 × 106 for the year 2020. The ESV lost between 1986 and 2020 was about US$ 92.11 × 106. In terms of the individual ecosystem services, the contribution of all service values declined except food production and biological control. Among the major contributors of ESVs, food production, water regulation, erosion control, climate regulation, nutrient cycling, waste treatment, and water supply represent 89%, 88%, 86% in 1986, 2003, and 2020, respectively. Soil erosion and sediment transport under environmental pressure significantly contributed for the loss of ecosystem services (ES) in the sub-basin. In addition, FGD and KII ranked provisioning ecosystem services first in order of importance for communities in the study area. Therefore, it is important to pay attention to the sustainable conservation and restoration of natural resources through natural vegetation protection, enhanced vegetation cover, conservation practices and wetland management to enhance the natural capital and ecosystem services in the Muger Sub-basin.

Divergent dynamics of coastal wetlands in the world’s major river deltas during 1990–2019

Coastal wetlands provide vital dynamic ecosystem services. They have become increasingly important after being linked to several sustainable developmental goals, resulting in a focus on their protection, management, and restoration. Therefore, there is an increasing need to detect and compare coastal wetland spatiotemporal dynamics in deltas at a global scale. In this study, we mapped and characterized coastal wetland spatiotemporal patterns for 1990–2019 in the world’s major river deltas using pixel frequency algorithms and Landsat-4/5 (TM), −7 (ETM + ), −8 (OLI), and Sentinel-2 (MSI) time-series imagery obtained from Google Earth Engine (GEE). Our map had a high overall accuracy (91.84 %) for 2019. Tidal flats were primarily distributed in North America (∼6.87 %) and Asia (∼5.91 %), whereas salt marshes were more commonly found in North America (∼45.39 %) and South America (∼10.61 %). Mangroves are more common in South America (∼11.86 %) and Asia (∼5.83 %), primarily because of the Amazon River Delta and tropical and subtropical regions of Asia, which host several large river deltas. South America had the largest coastal delta wetland area (798,569 km2), followed by Asia (640,251 km2), North America (581,977 km2), Africa (181,977 km2), Europe (140,759 km2), and Oceania (15,915 km2). There was a minor difference in the distribution of wetland vegetation and tidal flats in Asian coastal deltas, and the wetland vegetation area in Asia was greater than that in tidal flats on other continents. We found that the coastal wetland areas increased during 1990–2001, decreased during 2001–2012, and steadily increased during 2012–2019. Our study provides a baseline for monitoring the area, status, and health of the coastal wetlands in these river deltas.

FV-DLV3+: a light-weight flooded vegetation extraction method with attention-based DeepLabv3+

ABSTRACT Climate change and human activities have led to a decline in wetland vegetation and sharp reduction of wetland area, which has caused considerable damage to the hydrological cycle and ecological functions of many wetlands. Flooded vegetation (FV) is an important wetland landscape, and strengthening the research on methods of the extracting FV using remote sensing or aerial photos plays an important role in monitoring changes to wetlands. This paper used optical satellite and synthetic aperture radar satellite images as data sources, construct a set of feature variable sets suitable for FV extraction; in addition, a deep learning method based on the semantic segmentation model DeepLabv3+, FV-DLV3+, is proposed for FV extraction. The method adopts the lightweight MobileNetV2 model as the backbone feature extraction network, introduces the SegFormer network to make up for the feature information loss caused by MobileNetV2, and further captures more effective features and suppresses the interference of the background through an attention mechanism to enhance the ability to extract feature boundaries, thereby improving the accuracy of FV extraction. Taking the wetland of Poyang Lake as the experimental area, Landsat 8 and Sentinel-1A satellite data were selected as the data sources to produce a deep learning dataset, and U-Net, DeepLabv3+, Swin-Unet, Swin Transformer, and CVTNet were selected for comparison when conducting model training and validating experiments. The results show that FV-DLV3+ has an excellent effect on FV extraction of the Poyang Lake wetland, which effectively reduces the computational complexity of the process while improving extraction accuracy; the mean intersection over union of this paper’s method was higher than the five above-mentioned methods, which was 12.93% and 5.85% higher than that of DeepLabv3+ and CVTNet, respectively.

Differential impacts of fire and inundation on a wetland plant community after wildfire

Context Understanding fire and inundation impacts on wetland vegetation communities is crucial for effective post-fire wetland management. Aims We aimed to determine the impact of post-fire inundation on plant community structure and seedling germination and establishment after a large wildfire. We asked two questions, namely (1) did fire, drought or inundation affect plant communities the most; and (2) did fire or inundation affect seedling germination and establishment? Methods Using a before–after–control–impact (BACI) design, we monitored vegetation changes in water-couch wetland communities before and after a wildfire. Also, soil samples were collected from burnt and unburnt sites and assessed for impacts of fire and post-fire inundation regimes on seed germination and seedling establishment. Key results Inundation variables had more pronounced and more consistent impacts on vegetation measures than did fire or drought variables. Fire impacts were mainly short-term, with impact thresholds at 72 and 143 days after fire. Low germination levels and zero seedling survival were observed without inundation. Conclusions Fire was a major but short-term contributor to wetland vegetation change, whereas drought had longer-term impacts, and inundation regimes had the greatest impacts. Implications Providing a range of inundation conditions post-fire may enhance wetland vegetation recovery and have modifying effects on invasive species.

Impact of Extreme Drought on Vegetation Greenness in Poyang Lake Wetland

The Poyang Lake Wetland, an internationally significant ecosystem, frequently experiences drought during the flood season. However, the total impact of extreme drought on wetland vegetation remains poorly understood. This study determined the standardised precipitation evapotranspiration index (SPEI) and analysed drought trends within the Poyang Lake Basin. Additionally, spatiotemporal variations in wetland vegetation under drought conditions were examined by analysing the mean normalised difference vegetation index (NDVI) values and categorising NDVI classifications. The key factors affecting wetland vegetation and its respective thresholds were determined. The Poyang Lake Basin has experienced increasing aridity over the past 3 years. In response to this trend, the wetland vegetation area in Poyang Lake expanded, whereas vegetation greenness declined. Notably, in the year following an extreme drought, Poyang Lake’s vegetation greenness was lower than that during the same period in previous years. Regardless, the correlation analysis showed no significant relationship between the SPEI values and the wetland vegetation greenness; however, water level changes significantly impacted the wetland vegetation, with a correlation coefficient of −0.89 (p < 0.001). A critical water level of 14 m was identified as the threshold at which sudden changes in the mean NDVI were observed. This research offers valuable insights into hydrological management strategies to protect Poyang Lake Wetland’s vegetation under drought conditions. Future studies should enhance the differentiation of drought tolerance among different wetland plant species, thereby achieving differentiated hydrological management.

Devastation of the cerrado of mato grosso do sul and the advance of arenization in the pardo river watershed

Despite the significant biodiversity in the Cerrado, the process of occupation, subsequent economic cycles, and intensification of agriculture in a predatory manner have caused profound alterations in this biome. This study analyzed changes in land use and land cover in the Cerrado of Mato Grosso do Sul between 1985 and 2022 by mapping and also investigating the formation and dynamics of sand dunes in the Bauru sedimentary basin, focusing on the Pardo River watershed. MapBiomas Network data, collection 8, were used for land use and land cover analysis. Arenization foci cartography was conducted using brightness index 2 based on Sentinel-2 satellite images in Sentinel Application Platform software. The historical analysis of land use and cover in the Cerrado of Mato Grosso do Sul revealed a rapid transformation of the landscape, with a drastic reduction of native vegetation and an intensification of large-scale monocultures. Between 1985 and 2022, the Cerrado of Mato Grosso Sul lost 4.6 million hectares of native vegetation (forest formations, savannahs, grasslands, and wetlands), representing 24.54% of this territory. Recent changes in land use include pasture-soybean and pasture-eucalyptus. In the Pardo River watershed, arenization foci cover an area of 17,834.34 hectares, with varying dimensions ranging from less than 1 hectare to 376.41 hectares. The arenization process occurs in different slope compartments, such as the base, slope, and interfluves. The physical and chemical characteristics of Quartzipsamments, such as low organic carbon content in soil, are associated with high rates of surface exposure and weakened load-dependent structures, leading to degradation. Arenization occurs where natural vegetation has been suppressed and soil has been depleted by poorly managed production processes.

Coastal exotic plant serves as a habitat for a notorious wetland pest in unfavorable seasons: A case study of exotic Spartina alterniflora in China

Coastal biological invasions pose a wide-reaching threat to various ecosystems, affecting both vegetation and herbivores in native communities. Although herbivores often exert strong top-down control on vegetation, the impact of invasive species on consumers that strongly regulate native species in invaded ecosystems remains unclear. Therefore, through field surveys and feeding preference experiments, this study examined the effects of the invasive Spartina alterniflora Loisel. (hereafter, Spartina) on the native moth Laelia coenosa Hübner (hereafter, Laelia), a notorious pest that has been documented to cause significant damage to native Phragmites australis (Cav.) Trin. ex Steud. (hereafter, Phragmites) marshes in coastal wetlands of China. Field surveys showed that Laelia larvae were more abundant and had higher grazing rates in Phragmites than in Spartina marshes in summer, but these patterns reversed in autumn. Feeding experiments consistently showed that the dietary preference of Laelia larvae switched from Phragmites in summer to Spartina in autumn, likely because Spartina has a longer growing season and relatively higher nutritional quality than Phragmites in autumn. Thus, by providing shelters (i.e., dietary sources and habitats) during unfavorable seasons, Spartina invasions may facilitate this insect pest Laelia and potentially amplify its damage to native wetland vegetation. Our work reveals a novel, indirect mechanism of coastal invasion impacts and highlights the importance of incorporating seasonal variation in plant-herbivore interactions.

Reconstructing the rapid transitions of ecosystems during the mid-late Holocene: A pollen record from Haixing wetland in Bohai Bay, North China

Vegetation in coastal regions is sensitive to climate and sea-level changes. However, currently there is still a lack of understanding about the stability of the Holocene coastal ecosystem and the rapid conversion processes of different coastal wetland ecosystems. In this study, the high-resolution vegetation succession history of the southwest coast of Bohai Bay during the mid-late Holocene was revealed, based on pollen analysis and REVEALS model, and the rate of vegetation change was estimated. Furthermore, the characteristics and mechanisms of different ecosystem changes were explored. The results indicate that there was a prolonged transition from shallow sea to lagoon before the formation of the Haixing wetland during the interval of 7500–4100 cal yr BP. The rate of watershed vegetation change increase significantly under the common influence of the 4.2 ka event and human activities, causing a regime shift in the mountain forest ecosystems and a decrease in landscape resilience. It was characterized by a substantial reduction in broad-leaved forests (from>30% to<5%), especially for Quercus (∼3%), which have not recovered to pre-event levels since the end of the 4.2 ka event. Since 3500 cal yr BP, the Haixing area was completely detached from the influence of the Bohai Sea, forming freshwater wetlands. The local vegetation rapidly shifted from alkali-tolerant communities that followed the sea retreat to freshwater marsh plant communities. During 2000–1100 cal yr BP, under the strong impact of human activities, the succession of forest and grassland communities in the basin became more frequent. The area of coastal salt marshes expanded, with salt-tolerant plant communities taking the absolute advantage. The wetland vegetation landscape became closer to the modern appearance. Overall, our study provides new evidence for understanding the rapid evolution of coastal ecosystems in East Asia influenced by a combination of climate, hydrology and humans. It will help guide the coastal regions in facing the challenges of future global climate changes.

Recovery dynamics of wetland vegetation along a hydrological gradient in an agriculturally used inland valley in Uganda

Recovery dynamics of wetland vegetation along a hydrological gradient in an agriculturally used inland valley in Uganda

Identification of High-Quality Vegetation Areas in Hubei Province Based on an Optimized Vegetation Health Index

This research proposes an optimized method for identifying high-quality vegetation areas, with a focus on forest ecosystems, using an improved Vegetation Health Index (VHI). The study introduces the Land Cover Vegetation Health Index (LCVHI), which integrates the Vegetation Condition Index (VCI) and the Temperature Condition Index (TCI) with land cover data. Utilizing MODIS (Moderate Resolution Imaging Spectroradiometer) satellite imagery and Google Earth Engine (GEE), the study assesses the impact of land cover changes on vegetation health, with particular attention to forested areas. The application of the LCVHI demonstrates that forests exhibit a VHI approximately 25% higher than that of croplands, and wetlands show an 18% higher index compared to grasslands. Analysis of data from 2012 to 2022 in Hubei Province, China, reveals an overall upward trend in vegetation health, highlighting the effectiveness of environmental protection and forest management measures. Different land cover types, including forests, wetlands, and grasslands, significantly impact vegetation health, with forests and wetlands contributing most positively. These findings provide important scientific evidence for regional and global ecological management strategies, supporting the development of forest conservation policies and sustainable land use practices. The research results offer valuable insights into the effective management of regional ecological dynamics.

Spatio-temporal Dynamics of Land Use/Cover Change and Associated Carbon Stocks in Kanyabaha Wetland in Rukiga District, Uganda

Wetlands play an important ecological function of sequestering atmospheric carbon dioxide and thereby moderating adverse impacts of climate change. It is therefore important to understand the dynamics of carbon stocks in wetland vegetation and soils. This study investigated the spatio-temporal dynamics of aboveground, belowground, and total carbon stocks in Kanyabaha Wetland, located in Rukiga District, Uganda, spanning from 1990 to 2021. Through field sampling and laboratory analysis, aboveground carbon stocks were assessed by harvesting vegetation biomass and converting it to carbon stock using established conversion factors. Soil samples collected at different depths (0-20cm, 20-50cm, 50-100cm) were analyzed for soil organic carbon content to determine belowground carbon stocks. The study reveals variable spatio-temporal patterns of carbon stocks across land use types, with papyrus-dominated areas exhibiting the highest aboveground carbon stocks (49.66 tC/ha), followed by small-scale farmlands (33.73 tC/ha) and tree plantations (23.01 tC/ha). Conversely, built-up areas exhibit the lowest carbon stocks (1.29 tC/ha). Temporal analysis reveals fluctuating patterns in carbon stocks, with increases observed in built-up areas and small-scale farmlands, and decreases in grasslands and tree plantations that could be due to changes in hydrological cycle. Belowground carbon stocks follow similar trends, with papyrus areas maintaining the highest stocks (39.96 tC/ha), particularly at deeper soil depths that exhibit the highest carbon accumulation due to its extensive network of papyrus rhizome. Changes in land use, especially reclamation of the wetlands for farming and settlements affected carbon capture and storage in the wetland ecosystem. These findings highlight the importance of targeted conservation of natural wetlands and sustainable land management strategies in the Kanyabaha Wetland catchment for enhanced carbon sequestration. Further, in depth studies in the variability of carbon stocks due to various eco-climatic factors and anthropogenic activities are necessary to support sustainable wetland land management practices in Uganda.

Computation of drag coefficient for real vegetation in wetlands

The flow characteristics in wetlands and vegetated channels are depend on the physical structure, density, and pattern of vegetation. Estimating average velocity in vegetated wetlands requires an accurate determination of the drag coefficient. The innovation of this research lies in calculating the drag coefficient while considering the pattern shape, plant flexural rigidity, and vegetation structure. Laboratory experiments were conducted in a rectangular flume using a parallel pattern of Eleocharis plants at three densities: low, medium, and high, with discharge rates of 18.2, 23.7, and 28.8 L/s, respectively. Comparative analysis revealed that the equation proposed by Kothyari et al. (2009) [23] is just suitable for determining the drag coefficient on rigid cylinders with a staggered pattern and it should be improved for real vegetation with different pattern. A comprehensive equation was developed for real wetland vegetation, incorporating a new pattern coefficient for pattern shape (ζpp) and correction factor (η) to consider plant flexural rigidity and vegetation structure. The results demonstrate that this equation accurately predicts the drag coefficient (RMSE = 0.127, MAE = 0.107, and R2 = 0.9059) in channels with real vegetation with parallel and staggered patterns.

Effects of the comprehensive elimination of Spartina alterniflora along China's coast on blue carbon and scenario prediction after ecological restoration

Salt marshes cover the largest area among the three types of traditional blue carbon ecosystems in China's coastal zone, with the introduced smooth cordgrass (Spartina alterniflora Loisel.) being dominant in these marshes. The effects of eradicating S. alterniflora nationwide and the subsequent ecological restoration on blue carbon are unclear. This paper evaluates the variation in blue carbon during the national S. alterniflora eradication campaign, which involves mechanical tillage from 2022 to 2025, and proposes three scenarios for blue carbon changes after native vegetation is reestablished by 2050. The results show that, in 2025, plant carbon stock and soil carbon stock will decrease by 1.38 Tg C and 1.21 Tg C, respectively, in the areas where S. alterniflora has been removed and managed. Although blue carbon is reduced in coastal wetlands in 2025, carbon stock is expected to increase in restored native vegetated wetlands by 2050. S. alterniflora is resilient and competitive, posing a high risk in secondary invasion. Scenario Ⅰ suggests that S. alterniflora marshes could almost recover to their original state from 2022, with 7.70 Tg C stored in plant and soil carbon stocks. Scenario Ⅱ involves native vegetated wetlands coexisting with invasive S. alterniflora marshlands, with a total carbon stock estimated at 7.15 Tg C, reflecting a decrease of 0.39 Tg C in soil carbon stock and by 0.16 Tg C in plant carbon stock. In Scenario Ⅲ, mudflats dominant and native vegetated habitats are reestablished only in suitable sites, with the total carbon stock estimated at 5.63 Tg C, a 26.9% decrease compared to 2022 levels. While eradicating invasive S. alterniflora and restoring native vegetation in China's coast enhance the ecosystem services, it reduces blue carbon stocks. Therefore, developing additional strategies to increase carbon storage in coastal wetlands is necessary.

Analysis of Tamarix chinensis Forest Characteristics, Salt Ion Distribution, and Non-Structural Carbohydrate Levels in the Yellow River Delta: A Spatial Study Based on Proximity to the Shoreline.

The distribution of vegetation in coastal wetlands is significantly influenced by soil properties. However, the mechanisms of how soil characteristics impact the physiological processes of Tamarix chinensis forests remain underexplored. This study examined changes in the soil physicochemical properties and structural attributes of natural T. chinensis forests in the Yellow River Delta with increasing distance from the shoreline. T. chinensis trees were classified into healthy, intermediate, and dying categories based on growth potential, and dynamic changes in salt ions and non-structural carbohydrates (NSCs) were investigated. Results indicated that increasing distance from the shoreline corresponded to decreased soil salinity and pH, and increased soil moisture. T. chinensis mortality rate decreased, while tree height and ground diameter increased with distance. Soil salt content was positively correlated with T. chinensis mortality, but negatively correlated with tree height and ground diameter. Trees with lower growth potential had higher Na+ but lower K+ and K+/Na+ ratio. Soil salt content was positively correlated with root and stem Na+, while soil moisture was positively correlated with leaf NSCs. These findings suggest that soil salt content and moisture significantly influence T. chinensis ion absorption and NSC accumulation, with sodium toxicity being a key factor in the spatial distribution of T. chinensis forests.

Assessing the ecological response plant and soil to the seawalls in the Laizhou bay coastal wetland, China

Coastal wetlands are extremely vulnerable to both marine damage and human activities. In order to protect these wetlands, many artificial seawalls have been constructed. However, studies are required to understand how coastal wetlands will evolve under the influence of artificial seawalls. Therefore, to understand this succession process of plants and their adaptation to habitats divided by seawalls, two different habitats inside and outside the seawalls were selected in Laizhou Bay, China. The results showed that there were 5 plant species outside the seawalls that were lower than the 13 species inside. Additionally, the dominant plant species were varied between the two habitats, with mostly annual herbs observed outside the seawalls and perennial shrubs inside. Soil salinity was higher outside the seawalls, which was the key impact factor of soil nutrient differences. The distribution of annual and perennial species may be constrained by spatial differences in soil stoichiometry. Therefore, the plants in coastal wetlands vary significantly at a small scale in response to the disturbance of artificial seawalls. The differences in soil and plants between the two habitats divided by the artificial seawalls provide a new insight for evaluating the artificial coastal projects. The only way to reduce the effects of seawalls on natural coastal wetland vegetation and ecosystem functions is to restore connectivity of tidal flow inside and outside the seawalls.

Interactions between vegetation, sedimentation and flood inundation levels in wetlands

Wetlands produce key ecosystem services to mitigate the impacts of peak flows caused by pluvial or fluvial floods or storm surges. Sediment floods were characterized by a peak flow flowing over a simulated wetland, populated with two natural species. Floods have been drawn as flows of height H, into waters of height h, where H > h. Peak flow along the flume passed through: peak flow adjustment; peak flow; drag-dominated peak flow; and gravity current regimes. For high inundation wetland levels, settling rates of coarse and fine sediment were similar during the peak flow regime. At larger distances, sedimentation decreased monotonically, with higher sedimentation of fine particles. For low inundation levels, the sedimentation rate during the drag-dominated peak flow regime was higher for coarse particles. Vegetation decreased the inundation level needed for enhancing sedimentation. Our study then adds practical knowledge at considering that the synergies between the vegetation and the inundation level may enhance wetland services such as the mitigation of pluvial, fluvial or coastal floodings.

Carbon Dioxide Fluxes from Peri-urban Tidal Flats in South China

Coastal ecosystems are recognized as important carbon reservoirs. However, assessments of coastal carbon dioxide (CO2) fluxes often neglect unvegetated tidal flats, particularly in peri-urban areas. This study investigated the gross primary production (GPP), net primary production (NPP), and total respiration (TR) of three tidal flats in Hong Kong, one of the largest cities in South China, during both summer and winter seasons to understand their CO2 fluxes. Results showed that GPP of three tidal flats was significantly higher in winter than in summer. However, no significant seasonal variations in TR were observed among the tidal flats. Structural equation modelling was used to examine the drivers of CO2 fluxes in the three tidal flats. The model showed that temperature and microphytobenthos abundance were identified as positive drivers of GPP, while sediment mud content had a positive effect on TR. The estimated daily NPP of these tidal flats ranged from -0.853 to 0.112 g C m−2 d−1, which is lower than the mean value reported for global vegetated coastal wetlands. Despite some seasonal and spatial variations, those peri-urban tidal flats may be considered as weak CO2 sources rather than CO2 sinks.

Prediction of Sea Level Using Double Data Decomposition and Hybrid Deep Learning Model for Northern Territory, Australia

Sea level rise (SLR) attributed to the melting of ice caps and thermal expansion of seawater is of great global significance to vast populations of people residing along the world’s coastlines. The extent of SLR’s impact on physical coastal areas is determined by multiple factors such as geographical location, coastal structure, wetland vegetation and related oceanic changes. For coastal communities at risk of inundation and coastal erosion due to SLR, the modelling and projection of future sea levels can provide the information necessary to prepare and adapt to gradual sea level rise over several years. In the following study, a new model for predicting future sea levels is presented, which focusses on two tide gauge locations (Darwin and Milner Bay) in the Northern Territory (NT), Australia. Historical data from the Australian Bureau of Meteorology (BOM) from 1990 to 2022 are used for data training and prediction using artificial intelligence models and computation of mean sea level (MSL) linear projection. The study employs a new double data decomposition approach using Multivariate Variational Mode Decomposition (MVMD) and Successive Variational Mode Decomposition (SVMD) with dimensionality reduction techniques of Principal Component Analysis (PCA) for data modelling using four artificial intelligence models (Support Vector Regression (SVR), Adaptive Boosting Regressor (AdaBoost), Multilayer Perceptron (MLP), and Convolutional Neural Network–Bidirectional Gated Recurrent Unit (CNN-BiGRU). It proposes a deep learning hybrid CNN-BiGRU model for sea level prediction, which is benchmarked by SVR, AdaBoost, and MLP. MVMD-SVMD-CNN-BiGRU hybrid models achieved the highest performance values of 0.9979 (d), 0.996 (NS), 0.9409 (L); and 0.998 (d), 0.9959 (NS), 0.9413 (L) for Milner Bay and Darwin, respectively. It also attained the lowest error values of 0.1016 (RMSE), 0.0782 (MABE), 2.3699 (RRMSE), and 2.4123 (MAPE) for Darwin and 0.0248 (RMSE), 0.0189 (MABE), 1.9901 (RRMSE), and 1.7486 (MAPE) for Milner Bay. The mean sea level (MSL) trend analysis showed a rise of 6.1 ± 1.1 mm and 5.6 ± 1.5 mm for Darwin and Milner Bay, respectively, from 1990 to 2022.

Effects of disturbances on the spatiotemporal patterns and dynamics of coastal wetland vegetation

The coastal wetlands represented by mangroves are vulnerable to disturbances. To clarify the effects of disturbances quantitatively on mangroves, spatiotemporal changes in land use and land cover (LULC) were compared between the inside and outside of mangroves during 1988 and 2020 in Bangladesh by remote sensing. The mangrove region, which had diverse surface greenness due to natural and anthropogenic disturbances, were surveyed. Changes in greenness were estimated by five vegetation indices (VIs). To increase in the accuracy of land use classification, the five VIs were combined under a random forest (RF) classifier. The five VIs were: normalized difference vegetation index (NDVI), green NDVI (GNDVI), soil-adjusted VI (SAVI), green–red VI (GRVI) and enhanced VI (EVI), using data obtained from Landsat TM (30 m resolution) and Sentinel-2A (10 m) via Google Earth Engine. A Moran’s I analysis showed a random pattern of training samples of LULC classes for integrated RF classification. The classification accuracy was evaluated using the reference point data with multiple comparisons of VIs and κ coefficient. The area was classified into eight LULC classes. Although the respective VIs depicted unique characteristics, the accuracies of them were less than 89 %. The integrated classifier improved the accuracy of 96–97 % with κ of 0.96 in 1988 and 0.97 in 2020. The VI-integrated RF was the most effective classifier for LULC, particularly for separating mangroves from homestead vegetation. The changes for 32 years showed that the dense mangrove occupied 42 % of surveyed area in 1988 but declined ca 622 km2 in 2020, whereas sparse mangrove and aquaculture increased by 487 km2 and 464 km2, respectively, out of the total area of 8,360 km2. The changes in LULC classes were mostly from dense mangroves to sparse mangroves, sparse mangroves to dense mangroves and from cropland to bare land and aquaculture. Inaccessible mangrove degradation should result from natural disturbances, i.e., tropical cyclones, while areas near forest boundaries were distorted by anthropogenic disturbances such as aquaculture development. The restoration of habitat quality, mangrove afforestation and sustainable aquaculture practices were effective indicators for improving the mangrove ecosystems. Since natural and anthropogenic disturbances induce substantial loss of mangrove vegetation, the VI-integrated RF classifier should be applied for monitoring LULC and for developing the sustainable management strategies.