Special Issue "Inland Surface Water and Deep Learning"

A special issue of Water (ISSN 2073-4441). This special issue belongs to the section "New Sensors, New Technologies and Machine Learning in Water Sciences".

Deadline for manuscript submissions: closed (30 December 2022) | Viewed by 7959

Special Issue Editors

Prof. Dr. Min Feng
E-Mail Website
Guest Editor
Institute of Tibetan Plateau Research, Chinese Academy of Sciences, Chaoyang District, Beijing, 100101, China
Interests: Calibration and fusion of multi-sources; multi-types of remote sensing data; Long term forest and water monitoring and analysis; Periglacial landforms identification using Deep learning
Prof. Dr. Chengquan Huang
E-Mail Website
Guest Editor
Department of Geographical Sciences, University of Maryland, College Park, MD 20742, USA
Interests: land cover and land use change; vegetation structure and dynamics; disturbance, habitat, and carbon
Dr. Do-Hyung Kim
E-Mail Website
Guest Editor
Office of Innovation, United Nations Children’s Fund, New York, NY, USA
Interests: remote sensing; land use; land cover change; object detection; geospatial analysis
Special Issues, Collections and Topics in MDPI journals
Dr. Wenlong Jing
E-Mail Website
Guest Editor
Guangzhou Institute of Geography, Guangdong Academy of Sciences, Guangzhou 510070, China
Interests: remote sensing; water resources; machine learning; data fusion
Special Issues, Collections and Topics in MDPI journals

Special Issue Information

Dear Colleagues,

Inland surface water occupies only a small fraction of global land surface, but it plays a critical role in the sustainability of human society and terrestrial ecosystems. Many surface waters are small but providing critical ecosystem services (e.g., nutrient cycling, primary production, water provisioning, water purification, and recreation) to the ecosystems upon which numerous plant and animal species depend. Meanwhile, the dynamics of surface waters are generally faster and more dramatic than that of other land covers, posing great risks to its surroundings and downstream environments. The surface waters may also be at particular risk due to pressures such as unsustainable consumption, wetland drainage, land use intensification, stream diversion as well as climate change, especially in dry regions where water scarcity is becoming a major limiting factor for wildlife and humans. However, there are still noticeable gaps in our knowledge of inland surface water mainly due to the lack of effective approaches for mapping, identifying, and monitoring inland surface water.

In the past decade, Earth observations have been quickly accumulating by satellite, airborne, and UAV-based instruments, providing information with unprecedented precisions at spatial and temporal dimensions. This information provides a new foundation for a better understanding of the inland water around us. However, it is becoming challenging for traditional machine learning approaches to effectively extract thus information from those remotely sensed data with such large volume, high variety, and fidelity. The Deep Learning algorithms have been developed with the support of abundant data and powerful computing infrastructure, representing one of the most creative branches of machine learning and computer sciences. With its sophisticated architecture, Deep Learning has demonstrated superior intelligence on a wide range of applications in both industry and academy. Recently, there has been a surge of Deep Learning practices in remote sensing research on many topics, including those related to inland water.

We believe that Deep Learning would play a critical role in the future of inland surface water research in the upcoming big data era. We are organizing a special issue in Water to recognize the standing of Deep Learning and promote its application in water related research. For all the aforementioned, we kindly invite the community to contribute to this special issue, by submitting papers presenting deep learning related research of and beyond topics below:

  • Mapping Inland water using high-resolution remote sensing data.
  • Identifying types of water bodies, e.g., natural and artificial lakes, fishing ponds, glacier lakes.
  • Estimation of water quality and water pollutions.
  • Identification and assessment of water related disasters.
  • Developing deep learning algorithms for water.

Prof. Dr. Min Feng
Prof. Dr. Chengquan Huang
Dr. Dohyung Kim
Dr. Wenlong Jing
Guest Editors

Manuscript Submission Information

Manuscripts should be submitted online at www.mdpi.com by registering and logging in to this website. Once you are registered, click here to go to the submission form. Manuscripts can be submitted until the deadline. All submissions that pass pre-check are peer-reviewed. Accepted papers will be published continuously in the journal (as soon as accepted) and will be listed together on the special issue website. Research articles, review articles as well as short communications are invited. For planned papers, a title and short abstract (about 100 words) can be sent to the Editorial Office for announcement on this website.

Submitted manuscripts should not have been published previously, nor be under consideration for publication elsewhere (except conference proceedings papers). All manuscripts are thoroughly refereed through a single-blind peer-review process. A guide for authors and other relevant information for submission of manuscripts is available on the Instructions for Authors page. Water is an international peer-reviewed open access semimonthly journal published by MDPI.

Please visit the Instructions for Authors page before submitting a manuscript. The Article Processing Charge (APC) for publication in this open access journal is 2200 CHF (Swiss Francs). Submitted papers should be well formatted and use good English. Authors may use MDPI's English editing service prior to publication or during author revisions.

Published Papers (7 papers)

Order results
Result details
Select all
Export citation of selected articles as:

Research

Article
Identifying Alpine Lakes in the Eastern Himalayas Using Deep Learning
Water 2023, 15(2), 229; https://doi.org/10.3390/w15020229 - 05 Jan 2023
Viewed by 484
Abstract
Alpine lakes, which include glacial and nonglacial lakes, are widely distributed in high mountain areas and are sensitive to climate and environmental changes. Remote sensing is an effective tool for identifying alpine lakes over large regions, but in the case of small lakes, [...] Read more.
Alpine lakes, which include glacial and nonglacial lakes, are widely distributed in high mountain areas and are sensitive to climate and environmental changes. Remote sensing is an effective tool for identifying alpine lakes over large regions, but in the case of small lakes, the complex terrain and extreme weather make their accurate identification extremely challenging. This paper presents an automated method for alpine lake identification developed by leveraging deep learning algorithms and multi-source high-resolution satellite data. The method is able to detect the outlines and types of alpine lakes from high-resolution optical and Synthetic Aperture Radar (SAR) satellite data. In this study, a total of 4584 alpine lakes (including 2795 glacial lakes) were identified in the Eastern Himalayas from Sentinel-1 and Sentinel-2 data acquired during 2016–2020. The average area of the lakes was 0.038 km2, and the average elevation was 4974 m. High accuracy was reported for the dataset for both segmentation (mean Intersection Over Union (MIoU) > 72%) and classification (Overall Accuracy, User’s and Producer’s Accuracies, and F1-Score are all higher than 85%). A higher accuracy was found for the combination of optical and SAR data than relying on single-sourced data, for which the MIoU increased by at least 12%, suggesting that the combination of optical and SAR data is critical for improving the identification of alpine lakes. The deep learning-based method demonstrated a significant improvement over traditional spectral extraction methods. Full article
(This article belongs to the Special Issue Inland Surface Water and Deep Learning)
Show Figures

Figure 1

Article
A Comparison of Different Water Indices and Band Downscaling Methods for Water Bodies Mapping from Sentinel-2 Imagery at 10-M Resolution
Water 2022, 14(17), 2696; https://doi.org/10.3390/w14172696 - 30 Aug 2022
Cited by 1 | Viewed by 738
Abstract
Satellite-based remote sensing is important for monitoring the spatial distribution of water resources. The water index is currently one of the most widely used water body extraction methods. Based on Sentinel-2 remote sensing image, this study combines area-to-point regression kriging interpolation, bilinear interpolation, [...] Read more.
Satellite-based remote sensing is important for monitoring the spatial distribution of water resources. The water index is currently one of the most widely used water body extraction methods. Based on Sentinel-2 remote sensing image, this study combines area-to-point regression kriging interpolation, bilinear interpolation, and the Gram–Schmidt (GS) pan-sharpening method with the water indices MNDWI, AWEIsh and WI2015 to compare different water body extraction methods. The experimental results showed that all water indices have satisfactory extraction ability, with the kappa coefficient as an accuracy threshold above 0.8. Moreover, the GS downscaling method combined with the WI2015 yielded the best performance. This research demonstrates the efficacy of the WI2015 method to extract water bodies in urban areas and its ability to comprehensively describe river water bodies. The findings indicate that high-resolution band information is particularly important for improving low-resolution band downscaling results and can significantly minimize erroneous water body extraction. Full article
(This article belongs to the Special Issue Inland Surface Water and Deep Learning)
Show Figures

Figure 1

Article
Cyclicities in the Regime of Groundwater and of Meteorological Factors in the Basin of the Southern Bug River
Water 2022, 14(14), 2228; https://doi.org/10.3390/w14142228 - 15 Jul 2022
Cited by 1 | Viewed by 935
Abstract
The data of observations since 1951 in the upper part of the Southern Bug River basin in the west of Ukraine are analyzed. The results indicate that the climate change occurring on Earth disrupts the regular cyclicity of groundwater flow indicators. The identified [...] Read more.
The data of observations since 1951 in the upper part of the Southern Bug River basin in the west of Ukraine are analyzed. The results indicate that the climate change occurring on Earth disrupts the regular cyclicity of groundwater flow indicators. The identified 7–8-year groundwater level and flow to the river cyclicity correlates well with the air temperature, precipitation and river runoff cyclicity. The noted groundwater cyclicity appears with some delay after the establishment of the 8-year air temperature cyclicity observed since 1969. The manifestation of a 7–8-year cycle depends on the groundwater table (GWT) depth. For shallow groundwater (1.0–2.5 m), such rhythms have been observed since 1975, and for deeper levels, since 1989, which is recognized as the year of the beginning of the climate changes. Moreover, 7–8-year rhythms in the fluctuation of groundwater parameters are characteristic of mainly high-water periods of their multiyear regime, and during the low-flow phase is significantly weakened. During 2011–2014, the groundwater levels abnormally decreased and the 8-year cycles were replaced with 5-year ones. The influence of air temperature on the groundwater regime exceeds the role of other factors. Wavelet analysis was used as the main method of periodicity observation. Gaussian and Morlet wavelets provide the visualization of pronounced periodicities of data. Using multiple correlation analysis, it was confirmed that temperature has become the dominant impact factor on the groundwater (GWT 1.5–4.0 m) regime in recent decades. Full article
(This article belongs to the Special Issue Inland Surface Water and Deep Learning)
Show Figures

Graphical abstract

Article
Water Body Mapping Using Long Time Series Sentinel-1 SAR Data in Poyang Lake
Water 2022, 14(12), 1902; https://doi.org/10.3390/w14121902 - 13 Jun 2022
Cited by 2 | Viewed by 1170
Abstract
Mapping water bodies with a high accuracy is necessary for water resource assessment, and mapping them rapidly is necessary for flood monitoring. Poyang Lake is the largest freshwater lake in China, and its wetland is one of the most important in the world. [...] Read more.
Mapping water bodies with a high accuracy is necessary for water resource assessment, and mapping them rapidly is necessary for flood monitoring. Poyang Lake is the largest freshwater lake in China, and its wetland is one of the most important in the world. Poyang Lake is affected by floods from the Yangtze River basin every year, and the fluctuation of the water area and water level directly or indirectly affects the ecological environment of Poyang Lake. Synthetic Aperture Radar (SAR) is particularly suitable for large-scale water body mapping, as SAR allows data acquisition regardless of illumination and weather conditions. The two-satellite Sentinel-1 constellation, providing C-Band SAR data, passes over the Poyang Lake about five times a month. With its high temporal-spatial resolution, the Sentinel-1 SAR data can be used to accurately monitor the water body. After acquiring all the Sentinel-1 (1A and 1B) SAR data, to ensure the consistency of data processing, we propose the use of a Python and SeNtinel Application Platform (SNAP)-based engine (SARProcMod) to process the data and construct a Poyang Lake Sentinel-1 SAR dataset with a 10 m resolution. To extract water body information from Sentinel-1 SAR data, we propose an automatic classification engine based on a modified U-Net convolutional neural network (WaterUNet), which classifies all data using artificial sample datasets with a high validation accuracy. The results show that the maximum and minimum water areas in our study area were 2714.08 km2 on 20 July 2020, and 634.44 km2 on 4 January 2020. Compared to the water level data from the Poyang gauging station, the water area was highly correlated with the water level, with the correlation coefficient being up to 0.92 and the R2 from quadratic polynomial fitting up to 0.88; thus, the resulting relationship results can be used to estimate the water area or water level of Poyang Lake. According to the results, we can conclude that Sentinel-1 SAR and WaterUNet are very suitable for water body monitoring as well as emergency flood mapping. Full article
(This article belongs to the Special Issue Inland Surface Water and Deep Learning)
Show Figures

Figure 1

Article
Spatial Differentiation of Pond Landscapes across an Urban-Rural Gradient in the Pearl River Delta Region
Water 2022, 14(10), 1637; https://doi.org/10.3390/w14101637 - 20 May 2022
Viewed by 937
Abstract
The impact of ponds by urbanization has been widely discussed on the landscape scale. However, relatively few studies have explored the spatial differentiation of pond landscapes across urban-rural gradients in rapidly urbanising areas. In this study, we applied the DeepLabv3+ network to perform [...] Read more.
The impact of ponds by urbanization has been widely discussed on the landscape scale. However, relatively few studies have explored the spatial differentiation of pond landscapes across urban-rural gradients in rapidly urbanising areas. In this study, we applied the DeepLabv3+ network to perform a semantic segmentation on Google Map images to extract ponds in the Pearl River Delta (PRD) region, China; then we employed geographic information systems to analyse the ponds changes in 665 towns along an urban-rural gradient in the PRD. Results indicate that there are clear differences in landscapes between the urban core, urban, peri-urban, agricultural, and forested zone in terms of pond area and size. In total, 57.84% and 31.33% of the ponds are distributed in the peri-urban and agricultural zone, respectively; fewer ponds are present at either end of the urban-rural gradient. Owing to the legacy effects of historical land use and river systems, urban and peri-urban zone in the central and western parts of the PRD are still rich in ponds. We propose that management measures should be implemented according to the characteristics of different pond landscapes. Full article
(This article belongs to the Special Issue Inland Surface Water and Deep Learning)
Show Figures

Figure 1

Article
Deep Learning for Streamflow Regionalization for Ungauged Basins: Application of Long-Short-Term-Memory Cells in Semiarid Regions
Water 2022, 14(9), 1318; https://doi.org/10.3390/w14091318 - 19 Apr 2022
Cited by 3 | Viewed by 1011
Abstract
Rainfall-runoff modeling in ungauged basins continues to be a great hydrological research challenge. A novel approach is the Long-Short-Term-Memory neural network (LSTM) from the Deep Learning toolbox, which few works have addressed its use for rainfall-runoff regionalization. This work aims to discuss the [...] Read more.
Rainfall-runoff modeling in ungauged basins continues to be a great hydrological research challenge. A novel approach is the Long-Short-Term-Memory neural network (LSTM) from the Deep Learning toolbox, which few works have addressed its use for rainfall-runoff regionalization. This work aims to discuss the application of LSTM as a regional method against traditional neural network (FFNN) and conceptual models in a practical framework with adverse conditions: reduced data availability, shallow soil catchments with semiarid climate, and monthly time step. For this, the watersheds chosen were located on State of Ceará, Northeast Brazil. For streamflow regionalization, both LSTM and FFNN were better than the hydrological model used as benchmark, however, the FFNN were quite superior. The neural network methods also showed the ability to aggregate process understanding from different watersheds as the performance of the neural networks trained with the regionalization data were better with the neural networks trained for single catchments. Full article
(This article belongs to the Special Issue Inland Surface Water and Deep Learning)
Show Figures

Figure 1

Article
Impacts of Land-Use Change, Associated Land-Use Area and Runoff on Watershed Sediment Yield: Implications from the Kaduna Watershed
Water 2022, 14(3), 325; https://doi.org/10.3390/w14030325 - 22 Jan 2022
Cited by 3 | Viewed by 2061
Abstract
An uncontrolled sediment influx from the watershed upstream is a known threat to dam stability, while the pattern and amount of sediment yield are influenced by the predominant upstream land-use and land cover (LULC) types, precipitation amount, and intensity. Hence, the need to [...] Read more.
An uncontrolled sediment influx from the watershed upstream is a known threat to dam stability, while the pattern and amount of sediment yield are influenced by the predominant upstream land-use and land cover (LULC) types, precipitation amount, and intensity. Hence, the need to monitor sediment yield accumulation and its controlling factors in dam operation becomes crucial. In this paper, the Soil and Water Assessment Tool (SWAT) was used to assess the roles of land-use change, land cover area, and runoff on watershed’s sediment yield based on change detection analysis between 1975 and 2013 in the Kaduna Watershed (Nigeria), Western Africa. The SWAT standard procedures for the simulation of hydrological characteristics and sediment yields prediction were adopted. The datasets were calibrated for a period of 46 years and validated using 2015–2017 measured flow data, and suspended sediments concentration (SSC) acquired between March and October 2018. The model function was statistically determined using the Nash-Sutcliffe (NS), the coefficient of determination (r2) and the percentage of observed data (p-factor). The evaluation results of the SWAT model yielded NS, r2 and p-factor of 0.71, 0.80, and 0.86, respectively. These data suggest that the model performed satisfactorily for streamflow and sediment yield predictions. Findings suggest that the extinction of evergreen forests and a significant change in land-use from range grasses and forest to agriculture generic and residential types between 1975 and 2013, which resulted in surface runoff, sediment yield, and flow alteration. Evapotranspiration increased by 22.40% between 1975 and 2013. These changes have negatively impacted the watershed runoff by 56.00% and model sediment yield by 68.00% at the end of 2013. Thus, these variations can influence various human activities in the watershed, such as food security, livestock, energy production and water supply. It is hypothesized from the presented data that land use types exact a more dominant control on runoff and sediment yield than land cover area, although climatic influence may not be ruled out. Full article
(This article belongs to the Special Issue Inland Surface Water and Deep Learning)
Show Figures

Figure 1

Back to TopTop