Machine learning (ML) methods such as Random Forest (RF) have shown promises to estimate Secchi Disk Depth (Zsd). However, lack of a comprehensive dataset has been a long-lasting issue for training ML models in remote sensing of water quality. To aid the training process, the GLORIA dataset has recently provided access to hyperspectral in-situ measurements of remote sensing reflectance (Rrs) along with associated water quality parameters for globally representative inland and coastal waters. We use simulated Sentinel-2 Rrs to train a global model using GLORIA and then validate it on independent data from Finger Lakes, USA. When compared to RF model trained on Finger Lakes data, the validation results indicate better performance (Mean Absolute Error (MAE) 37%) as compared to the global model trained on GLORIA (MAE 94%). However, when the global model was validated on independent dataset from GLORIA (i.e. Lake Erie), the results were promising (MAE 34%). Therefore, the models can be used to estimate Zsd globally, provided the uncertainties in deriving satellite based Rrs are accounted for.

Forest ecosystems have been persistently affected by wildfires, leading to significant damage worldwide. The severity and frequency of wildfires have escalated in recent years, necessitating more effective prediction models. This study presents an application of convolutional neural networks (CNNs) for wildfire spread prediction, focusing on the use of atrous spatial pyramid pooling (ASPP) mechanisms in these networks. However, the black-box nature of these algorithms has not been fully explored. To bridge this gap, we proposed an explainable CNN model with an ASPP mechanism (CNN-ASPP) in this study. More specifically, we utilize the Next Day Wildfire Spread dataset, which includes environmental variables, to evaluate the performance of our model. The proposed model is compared with state-of-the-art machine learning (ML) methods, including random forest (RF), support vector machine (SVM), artificial neural network (ANN), and another CNN model. Our results showed that CNN-ASPP achieved an F1-score of 97%, outperforming the ML methods with an F1-score of 90% for a neighborhood size of$7\times 7$. We also opened the black box and tried to explanation different convolutional layers based on the gradient-weighted class activation mapping (Grad-CAM) algorithm. Our findings indicate that larger dilation rates (DRs) can extract more meaningful features from the input data. This study contributes to the development of more transparent and accurate models for wildfire spread prediction, which could have significant implications for forest management and wildfire prevention strategies.

Effective monitoring of wetlands plays a pivotal role in comprehending and managing these ecologically vital ecosystems. This study assesses the potential of C-band synthetic aperture radar (SAR) imagery in compact polarization (CP) mode, utilizing the RADARSAT Constellation Mission (RCM), for wetland characterization. We introduce the compact-polarimetric signature (CPS) as a novel descriptor to delineate wetlands, including bog, fen, and marsh classes. In addition, we propose an alternative decomposition technique ($\mu -\chi $) to segment the total power into three components: odd-bounce scattering$(P_{s})$, double-bounce scattering$(P_{d})$, and random scattering$(P_{v})$. For our evaluation, we selected a test site in New Brunswick, Canada, and acquired a series of RCM datasets covering this region. The time-series CPS plots yield valuable insights, elucidating the scattering mechanisms of different wetland classes. Notably, these plots reveal that during the active season, characterized by changing vegetation structures, the scattered waves exhibit variations, leading to changes in received power and the purity parameter ($\mu $). Furthermore, the observed variations in the proposed power components demonstrate a significant discriminatory capacity among wetlands. The$P_{s}$,$P_{d}$, and$P_{v}$components effectively distinguish bog, fen, and marsh classes, respectively, capturing the unique characteristics of each wetland type. These findings carry considerable potential for advancing wetland characterization through the RCM CP-SAR mission. The improved discriminative ability among different wetland classes is a valuable contribution to the broader field of wetland ecology and management. This advancement potentially empowers precise wetland classification, facilitating well-informed decision-making in wetland preservation and resource allocation. The applications of these findings extend to ecosystem monitoring, environmental impact assessments, and the long-term evaluation of wetland health. Eventually, this contributes to developing more effective wetland conservation and management strategies.

Wetlands serve critical functions for protecting ecosystems but are declining globally due to anthropogenic activities and environmental change. There are challenges in delineating wetlands that partially cannot be addressed using the integration of optical and SAR data. NASA’s global ecosystem dynamics investigation (GEDI) data, forest height products can differentiate wetland types with different heights. However, GEDI data is only available for limited footprints. In this study, we used a random forest regression model to create a canopy height model (CHM) by merging discrete GEDI footprints with Sentinel-2 data. We calculated the R-squared and RMSE using test data (0.81 and 3.48 meters, respectively). Such an investigation might provide the basis for better wetland management.

This study contributes to advancing the understanding and utilization of compact polarimetry (CP) RCM Synthetic Aperture Radar (SAR) data for enhanced crop characterization and mapping. The received wave polarization signature captures the explicit variation of the received power with a fixed transmit polarization and varying received polarization bases. This information is then suitably utilized for improved discrimination among multiple crop types. Furthermore, the study explores using multi-date polarimetric features extracted from RCM imagery to achieve more accurate and detailed crop mapping results. By incorporating information from multiple acquisition dates, the multi-date polarimetric features illustrate excellent potential in capturing temporal variations in crop characteristics, leading to enhanced crop mapping accuracy. The implications and findings from this study could be essential in demonstrating the role of RCM data in agricultural applications.

The growing effects of climate change and urbanization necessitates continuous monitoring of the freshwater resources in terms of water quality. Although remote sensing techniques have been successful in estimating water quality, its applicability over small oligotrophic lakes still remains a challenge due to the lower contribution of constituents to the water-leaving radiance. As such, this study leverages the availability of citizen science data and the synergistic use of Sentinel-2 imagery with Random Forest (RF) regression to estimate Secchi Disk Depth (SDD) over Canandaigua Lake. The results indicate an R2of 0.74, RMSE of about 0.72 m, MAE and Bias of 1.11 and 0.98, respectively. The feature importance for RF was also calculated, and the results indicate high value for visible bands. The model can be replicated for similar study areas and the findings can be used for efficient freshwater monitoring.

Investigating the quantity of forest above-ground biomass (AGB) is vital for understanding the role of forests in the global carbon cycle. Canopy height model (CHM) plays an important role in estimating AGB. Accurate large-scale forest CHM estimation using traditional methods requires a lot of labor, time, and cost. Remote sensing is a cost-effective approach, which provides valuable information over large areas in a timely manner. Recent advances in spaceborne light detection and ranging (LiDAR) data paved the road for measuring elevation. The global ecosystem dynamics investigation (GEDI) onboard the International Space Station is particularly designed to collect information on vertical vegetation structure. Thus, the main objective of this paper is to use the GEDI level 2A elevation and height metrics product to create a high resolution, state-wide CHM. To create a continuous CHM, GEDI point-based height measurements were extrapolated using Sentinel-2 imagery to produce a 10 m CHM of the New York State for the year 2019. The generated 10 m CHM was evaluated using GEDI height measurements (RMSE=4.85 m, R2=0.65). A comparison of 10 m CHM and a 30 m global CHM (RMSE=6.6 m, R2=0.62) demonstrated the potential of finer spatial resolution and red-edge bands in creating a more accurate CHM which is important to support forest monitoring at large-scale.

The Canadian RADARSAT Constellation Mission (RCM) has passed its early operation phase, with the current performance evaluation. In this study, we provide results update on RCM performance for selected SAR applications. The RCM was designed to address three core applications-disaster management, maritime surveillance, and ecosystem monitoring. Our study shows a promising level of agreement between RCM and RADARSAT-2 performance in flood mapping using dual-polarized HH-HV SAR data over Red River, Manitoba. Visual analysis of coincident RCM compact polarimetric and RADARSAT-2 dual-polarized HH-HV SAR imagery over the Resolute Passage, Canadian Central Arctic, highlighted an improved contrast between sea ice classes in dry ice winter conditions. Object-oriented classification of a wetland area in Newfoundland and Labrador by fusion of RCM dual-polarized VV-VH data and Sentinel-2 optical imagery revealed promising classification results, with an overall accuracy of 91.1% and a kappa coefficient of 0.87.

Wetlands provide different environmental services and contribute to global biodiversity and climate mitigation. However, these priceless environments have been damaged by direct and indirect interference of human activities. Monitoring wetland's hydroperiod over large study areas became feasible only by recent advances in cloud computing platforms and free archive of Landsat time-series. We used the Dynamic Surface Water Extent algorithm for the unsupervised classification of time-series images. This unsupervised algorithm showed 86.7% overall accuracy for 2020 results. Then hydroperiod maps were produced based on the water occurrence of each pixel, and a differencing change detection method was used to produce change maps.

In Canada, wetlands cover a large extent of landscape, providing necessary services to ecosystems. Wetland mapping and monitoring utilizing satellite Earth Observation (EO) imagery are of great importance for natural resource management. This study established a two-stream deep learning framework based on a Convolutional Neural Network (CNN) and Graph Convolutional Network (GCN) for wetland classification. The proposed architecture includes a dedicated information source for each stream, wherein Sentinel-1 and Sentinel-2 that are independently fed to GCN and CNN networks, respectively. The final classification result is achieved by concatenating the features extracted by each stream. The study area is a part of the Avalon site located on the Island of Newfoundland in Canada, where we have collected ground truth data for different wetland types. The result was compared to that obtained using Random Forest (RF) and Support Vector Machine (SVM) algorithms. The overall accuracies were 87.68%, 85.52%, and 83.7% for the proposed method, the RF, and the SVM, respectively.

Convolutional Neural Networks (CNNs) have shown promising results in classifying complex remote sensing scenery, particularly in the classification of wetlands. State-of-the-art Natural Language Processing (NLP) algorithms, on the other hand, are transformers. In this paper, we illustrate the effectiveness of the cutting-edge Swin Transformer for the classification of complex wetlands in New Brunswick, Canada. The precision of the proposed transformer is 0.66, 0.71, 0.75, 0.78, 0.82, 0.83, 0.84, 0.90, 0.90, 0.95, and 0.98 for the recognition of shrub, fen, forested wetland, crop, bog, freshwater marsh, coastal marsh, aquatic bed, grass, urban, and water, respectively. Based on the results, with a relatively high level of overall accuracy of slightly less than 80%, the proposed Swin Transformer is highly capable of complex wetland classification.

Monitoring freshwater quality is a global concern because of increasing harmful algal blooms (HABs). Therefore, it is important to detect HABs especially in small lakes as they hold great socioeconomic value. This study estimates the potential of using Sentinel-2 for estimating chlorophyll-a value in small inland lakes. In particular, this study uses support vector regression (SVR), random forest (RF) and adaptive boosting (AB) for Seneca lake. The processing power of Google Earth Engine (GEE) was used to extract the input features. The results indicate the superior performance of machine learning models in comparison to linear regression. Furthermore, AB provided the best results (R2=0.85) as compared to RF and SVR. In terms of ensemble, the combination of all three models performed best in terms of R2 (0.76), RMSE (0.633 µg/L) and MAE (0.728 µg/L).

Arctic environments are remote and inaccessible, making conventional field-based data collection challenging. Thus, this study describes an efficient desktop-based methodology for deriving reference data to support large-scale remote sensing classification focusing on wetland ecosystems. Our study area was Canada's Southern Arctic Ecozone. Various Earth observation (EO) datasets, including optical, multi-spectral, and topographic, were used as a base to support a photointerpretation process for collecting reference data. Ten 30-by-30-kilometer sampling plots were established across the ecozone for this activity based on a suite of minimum criteria. Reference polygons were assigned to one of the five major wetland classes of the Canadian Wetland Classification System (CWCS), along with a detailed wetland type definition. It is anticipated this methodology will be applied later to other northern ecozones to support large-scale wetland classification updates and status and trends reporting.

We propose a hybrid algorithm for despeckling the Synthetic Aperture Radar (SAR) images using the Convolutional Neural Network (CNN) denoising and complex wavelet shrinkage. In particular, we perform the speckle reduction process in the complex wavelet domain. We first despeckled the approximation complex wavelet coefficients using the MUltichannel LOgarithm with the Gaussian denoising algorithm (MuLoG) based on a pre-trained CNN model named FFDNet. Next, we despeckled the log-transformed details of the complex wavelet coefficients using the averaged version of the Maximum a Posteriori (AMAP) estimator. The experimental results on simulated and real SAR images showed that the proposed method achieved better speckle suppression in the homogeneous areas while preserving edges and point targets than other state-of-the-art methods.

Shallow convolutional neural networks (CNNs) have successfully been used to classify polarimetric synthetic aperture radar (PolSAR) imagery. However, one drawback of the existing deep CNN-based techniques is that the input PolSAR training data are often insufficient due to their need for a significant number of training data compared to shallow CNN models utilized in PolSAR image classification. In this paper, we propose using Haar wavelet transform in deep CNNs for effective feature extraction to improve the classification accuracy of PolSAR imagery. Based on the results, the proposed deep CNN model obtained better average accuracy in the San Francisco region with an accuracy of 93.3% and produced more homogeneous classification maps with less noise compared to the two much shallower CNN models of AlexNet (87.8%) and a 2D CNN network (91%). The proposed algorithm is efficient and may be applied over large areas to support regional wetland mapping and monitoring activities using PolSAR imagery. The codes are available at (https://github.com/aj1365/DeepCNN_Polsar).

While deep learning models have been extensively applied to land-use land-cover (LULC) problems, it is still a relatively new and emerging topic for separating and classifying wetland types. On the other hand, ensemble learning has demonstrated promising results in improving and boosting classification accuracy. Accordingly, this study aims to develop a classification system for mapping complex wetland areas by incorporating deep ensemble learning and satellite datasets. To this end, time series of Sentinel-1 dual-polarized Synthetic Aperture Radar (SAR) dataset, alongside Sentinel-2 multispectral imagery (MSI), are used as input data to the model. In order to increase the diversity of the extracted features, the proposed model, herein called WetNet, consists of three different submodels, comprising several recurrent and convolutional layers. Furthermore, multiple ensembling sections are added to different stages of the model to increase the transferability of the model (to other areas) and the reliability of the final results. WetNet is evaluated in a complex wetland area located in Newfoundland, Canada. Experimental results indicate that WetNet outperforms the state-of-the-art deep models (e.g., InceptionResnetV2, InceptionV3, and DenseNet121) in terms of both the classification accuracy and processing time. This makes WetNet an efficient model for large-scale wetland mapping application. The python code of the proposed WetNet model is available at the following link for the sake of reproducibility:https://colab.research.google.com/drive/1pvMOd3_tFYaMYGyHNfxqDxOiwF78lKgN?usp=sharing

Leaf Area Index (LAI) and biomass are the most critical biophysical parameters for crop monitoring. In this study, we used three ensemble-based methods, including Random Forest (RF), Gradient Boosting (GB), and Extreme Gradient Boosting (XGB), for crop parameter estimation and mapping of soybean and wheat in an agricultural region in Winnipeg, Canada. Various Vegetation Indices (VIs) and radar parameters were extracted from multitemporal multispectral Sentinel-2 images and Synthetic Aperture Radar (SAR) Sentinel-1 data. Feature selection was made, first, based on the correlation between extracted features and target biophysical parameters. Features with low importance were then removed based on the correlation between all features. The RF model has the lowest RMSE among the examined methods for dry biomass, wet biomass, and LAI for soybean. For wheat, XGB has the lowest RMSE for dry and wet biomasses, while RF led to LAI's highest accuracy.

Wetlands are significant global contributors to the environment and climate, and there are increasing efforts for wetland conservation globally. Recent advances in cloud computing platforms and accessibility of free medium resolution data lead to affordable solutions for large scale wetland mapping with remote sensing tools. Three Northern provinces of Iran include several complex wetland regions. A classification scheme consists of four wetland classes and five upland classes were chosen to describe wetland types for this region. A combination of Sentinel-2 surface reflectance summer composite and Sentinel-l synthetic aperture radar (SAR) datasets were used to train the machine learning model. Simple non-iterative clustering (SNIC) and Random Forest classification were implemented in Google Earth Engine (GEE) to produce an object-based wetland inventory map with an overall accuracy of 94.10%.

Wetlands are highly sensitive ecosystems that have experienced largely undocumented loss across Canada. Accurate statistics of historic loss of wetlands across many provinces is vague at best or non-existent at worst, as exemplified in Newfoundland and Labrador (NL). Thus, NL represents a perfect candidate for implementing historical remote sensing data sets and change detection methods. Given recent advancements in earth observation technology, it is now feasible to implement remote sensing-based change detection methods at scales never previously possible. As such, the goal of this work is to develop a methodology to assess wetland class change across the island of Newfoundland between 1985 and 2015 using historic and current Landsat imagery, Random Forest classification, and the Google Earth Engine (GEE) platform. The resulting accuracies ranged from 84.37% to 88.96%. The analysis reveals that wetland classes over the last 30 years have been unstable, and the biggest loss of wetlands to anthropogenic land cover occurred between the 1980's and the 1990's. Index Terms - Wetlands, Change Detection, Landsat, Geo big data

Shrub willow is a valuable source of hardwood biomass feedstock which is used for the production of bioenergy, biofuels, and renewable bio-based products. The biomass produced from this short-rotation woody plant can be used for heat and electricity generation. Thus, an accurate estimation of shrub willow above-ground biomass (AGB) is of paramount importance. This paper aimed to estimate shrub willow AGB using multi-spectral unmanned aerial system (UAS) imagery and machine learning techniques. To accomplish this goal, a machine learning model (i.e., random forest (RF)) and a deep learning method (i.e., convolutional neural network (CNN)) were applied to the spectral bands and some vegetation indices over a site in Camillus, NY, US in July 2019. The results demonstrated the superiority of the RF model (RMSE of 1.73 Mg/ha and R2 of 0.95) compared to the CNN (RMSE of 2.69 Mg/ha and R2 of 0.89) technique. Adding vegetation indices to spectral bands and using a convolutional approach for training purposes could significantly improve the modeling efficiency.

The synergistic use of spatial features with spectral properties of satellite images enhances thematic land cover information. This study aims to address the lack of high-level features by proposing a classification framework based on convolutional neural network (CNN) to learn deep spatial features for wetland. In particular, a CNN model was used for classification of remote sensing imagery with limited number of training data by fine-tuning of a preexisting CNN (AlexNet). The classification results obtained by the deep CNN were compared with those based on well-known ensemble classifiers, namely Random Forest (RF), to evaluate the efficiency of CNN Experimental results demonstrated that CNN was superior to RF for complex wetland mapping even by incorporating the small number of input features (i.e., 3 features) for CNN compared to RF. The proposed classification scheme serves as a baseline framework to facilitate further scientific research using the latest state-of-art machine learning tools for processing remote sensing data.

In this study, a new hierarchical object-based Random Forest (RF) classification approach is proposed for discriminating between different wetland classes in a study area located in the north eastern portion of the Avalon Peninsula, Newfoundland and Labrador province, Canada. Specifically, multi-polarization and multi-frequency SAR data, including single polarized TerraSAR-X (HH), dual polarized L-band ALOS-2 (HH/HV), and fully polarized C-band RADARSAT-2 images, were applied in three different classification levels. The overall accuracy and kappa coefficient were determined in each classification level for evaluating the classification results. Importantly, an overall accuracy of94.82% was obtained for the final classified map in this study.

An adaptive speckle reduction algorithm for Polarimetric SAR (PolSAR) data, based on the combination of Gaussian Markov Random Field (GMRF) and Wavelet Edge Detection (WED) is proposed in this paper. The algorithm has three major steps: (a) first-time speckle reduction based on the GMRF model, (b) detail preservation using a WED approach, and (c) second-time speckle reduction using a least square approach based on pseudo span image. Both the GMRF and WED use the coherency matrix as the input, which has sensitive diagonal elements, namely T11, T22and T33corresponding to surface, double-bounce, and volume scattering, respectively. A key point in the proposed algorithm is that strong point targets are not affected by speckle phenomena and thus, they should be excluded from the de-speckling process. The proposed algorithm is applied to a full polarimetric C-band RADARSAT-2 data in the Avalon Peninsula, Newfoundland and Labrador, Canada.

In this study, we evaluate the capability of Interferometric Synthetic Aperture Radar (InSAR) technique for the monitoring wetland water level changes in the Avalon Peninsula, Newfoundland and Labrador, Canada. This province is one of the richest Canadian provinces in terms of wetland expanse, yet these productive habitats remain poorly understood in this area. The InSAR technique is proven to be efficient in monitoring solid earth changes (e.g., earthquake and landslide). However, the use of such a technique for monitoring water level changes is underdeveloped and limited to particular pilot sites. In this paper, we use 5 SLC TerraSAR-X descending track and analyze them using repeat-pass SAR interferometry technique to monitor water level fluctuations of flooded vegetation.

Synthetic Aperture Radar (SAR) satellite sensors recently provide valuable sources of earth observation data for various environmental applications. Beside the specifics properties of these data including multi-polarization and polarimetric image data, the presence of unavoidable speckle seriously degrades the quality of these data. Specifically, in certain applications such as clustering, classification and change detection speckles make some difficulties in analysis data and interpretation of results. In this research, a hybrid approach, based on frequency-domain transforms, is proposed. This method is a combination of wavelet and curvelet transforms to suppress the speckle noise in SAR images. This approach based on features and region which has a good efficiency in removing noise and preserving information of data in case of edges and shape. Results of these methods were compared simultaneously and with conventional speckle filtering methods (e.g. Lee, Frost and Kuan).

This paper proposes a novel speckle reduction method that combines an advanced statistical distribution with spatial contextual information for SAR data. The method for despeckling is based on a Markov random field (MRF) that integrates a K-distribution for the SAR data statistics and a Gauss-MRF model for the spatial context. These two pieces of information are combined based on weighted summation of pixel-wise and contextual models. This not only preserves edge information in the image, but also improves signal-to-noise ratio (SNR) of the despeckled data. Experiments on real SAR data demonstrate the effectiveness of the algorithm compared with well-known despeckling methods.