Large-scale Land Cover Mapping Research Articles

Large-Scale Land Cover Mapping Framework Based on Prior Product Label Generation: A Case Study of Cambodia

Large-Scale land cover mapping (LLCM) based on deep learning models necessitates a substantial number of high-precision sample datasets. However, the limited availability of such datasets poses challenges in regularly updating land cover products. A commonly referenced method involves utilizing prior products (PPs) as labels to achieve up-to-date land cover mapping. Nonetheless, the accuracy of PPs at the regional level remains uncertain, and the Remote Sensing Image (RSI) corresponding to the product is not publicly accessible. Consequently, the sample dataset constructed through geographic location matching may lack precision. Errors in such datasets are not only due to inherent product discrepancies, and can also arise from temporal and scale disparities between the RSI and PPs. In order to solve the above problems, this paper proposes an LLCM framework for generating labels for use with PPs. The framework consists of three main parts. First, initial generation of labels, in which the collected PPs are integrated based on D-S evidence theory and initial labels are obtained using the generated trust map. Second, for dynamic label correction, a two-stage training method based on initial labels is adopted. The correction model is pretrained in the first stage, then the confidence probability (CP) correction module of the dynamic threshold value and NDVI correction module are introduced in the second stage. The initial labels are iteratively corrected while the model is trained using the joint correction loss, with the corrected labels obtained after training. Finally, the classification model is trained using the corrected labels. Using the proposed land cover mapping framework, this study used PPs to produce a 10 m spatial resolution land cover map of Cambodia in 2020. The overall accuracy of the land cover map was 91.68% and the Kappa value was 0.8808. Based on these results, the proposed mapping framework can effectively use PPs to update medium-resolution large-scale land cover datasets, and provides a powerful solution for label acquisition in LLCM projects.

Unsupervised Domain Adaptation with Contrastive Learning-Based Discriminative Feature Augmentation for RS Image Classification

High- and very high-resolution (HR, VHR) remote sensing (RS) images can provide comprehensive and intricate spatial information for land cover classification, which is particularly crucial when analyzing complex built-up environments. However, the application of HR and VHR images to large-scale and detailed land cover mapping is always constrained by the intricacy of land cover classification models, the exorbitant cost of collecting training samples, and geographical changes or acquisition conditions. To overcome this limitation, we propose an unsupervised domain adaptation (UDA) with contrastive learning-based discriminative feature augmentation (CLDFA) for RS image classification. In detail, our method first utilizes contrastive learning (CL) through a memory bank in order to memorize sample features and improve model performance, where the approach employs an end-to-end Siamese network and incorporates dynamic pseudo-label assignment and class-balancing strategies for adaptive domain joint learning. By transferring classification models trained on a source domain (SD) to an unlabeled target domain (TD), our proposed UDA method enables large-scale land cover mapping. We conducted experiments using a massive five billion-pixels dataset as the SD and tested the HR and VHR RS images of five typical Chinese cities as the TD and applied the method on the completely unlabeled world view 3 (WV3) image of Urumqi city. The experimental results demonstrate that our method excels in large-scale HR and VHR RS image classification tasks, highlighting the advantages of semantic segmentation based on end-to-end deep convolutional neural networks (DCNNs).

Improving Dryland Urban Land Cover Classification Accuracy Using a Classical Convolution Neural Network

Reliable information of land cover dynamics in dryland cities is crucial for understanding the anthropogenic impacts on fragile environments. However, reduced classification accuracy of dryland cities often occurs in global land cover data. Although many advanced classification techniques (i.e., convolutional neural networks (CNN)) have been intensively applied to classify urban land cover because of their excellent performance, specific classification models focusing on typical dryland cities are still scarce. This is mainly attributed to the similar features between urban and non-urban areas, as well as the insufficient training samples in this specific region. To fill this gap, this study trained a CNN model to improve the urban land classification accuracy for seven dryland cities based on rigorous training sample selection. The assessment showed that our proposed model performed with higher overall accuracy (92.63%) than several emerging land cover products, including Esri 2020 Land Cover (75.55%), GlobeLand30 (73.24%), GLC_FCS30-2020 (69.68%), ESA WorldCover2020 (64.38%), and FROM-GLC 2017v1 (61.13%). In addition, the classification accuracy of the dominant land types in the CNN-classified data exceeded the selected products. This encouraging finding demonstrates that our proposed architecture is a promising solution for improving dryland urban land classification accuracy and compensating the deficiency of large-scale land cover mapping.

Enabling country-scale land cover mapping with meter-resolution satellite imagery

High-resolution satellite images can provide abundant, detailed spatial information for land cover classification, which is particularly important for studying the complicated built environment. However, due to the complex land cover patterns, the costly training sample collections, and the severe distribution shifts of satellite imageries caused by, e.g., geographical differences or acquisition conditions, few studies have applied high-resolution images to land cover mapping in detailed categories at large scale. To fill this gap, we present a large-scale land cover dataset, Five-Billion-Pixels. It contains more than 5 billion labeled pixels of 150 high-resolution Gaofen-2 (4 m) satellite images, annotated in a 24-category system covering artificial-constructed, agricultural, and natural classes. In addition, we propose a deep-learning-based unsupervised domain adaptation approach that can transfer classification models trained on labeled dataset (referred to as the source domain) to unlabeled data (referred to as the target domain) for large-scale land cover mapping. Specifically, we introduce an end-to-end Siamese network employing dynamic pseudo-label assignment and class balancing strategy to perform adaptive domain joint learning. To validate the generalizability of our dataset and the proposed approach across different sensors and different geographical regions, we carry out land cover mapping on five megacities in China and six cities in other five Asian countries severally using: PlanetScope (3 m), Gaofen-1 (8 m), and Sentinel-2 (10 m) satellite images. Over a total study area of 60,000 km2, the experiments show promising results even though the input images are entirely unlabeled. The proposed approach, trained with the Five-Billion-Pixels dataset, enables high-quality and detailed land cover mapping across the whole country of China and some other Asian countries at meter-resolution.

DeepForest: Novel Deep Learning Models for Land Use and Land Cover Classification Using Multi-Temporal and -Modal Sentinel Data of the Amazon Basin

Land use and land cover (LULC) mapping is a powerful tool for monitoring large areas. For the Amazon rainforest, automated mapping is of critical importance, as land cover is changing rapidly due to forest degradation and deforestation. Several research groups have addressed this challenge by conducting local surveys and producing maps using freely available remote sensing data. However, automating the process of large-scale land cover mapping remains one of the biggest challenges in the remote sensing community. One issue when using supervised learning is the scarcity of labeled training data. One way to address this problem is to make use of already available maps produced with (semi-) automated classifiers. This is also known as weakly supervised learning. The present study aims to develop novel methods for automated LULC classification in the cloud-prone Amazon basin (Brazil) based on the labels from the MapBiomas project, which include twelve classes. We investigate different fusion techniques for multi-spectral Sentinel-2 data and synthetic aperture radar Sentinel-1 time-series from 2018. The newly designed deep learning architectures—DeepForest-1 and DeepForest-2—utilize spatiotemporal characteristics, as well as multi-scale representations of the data. In several data scenarios, the models are compared to state-of-the-art (SotA) models, such as U-Net and DeepLab. The proposed networks reach an overall accuracy of up to 75.0%, similar to the SotA models. However, the novel approaches outperform the SotA models with respect to underrepresented classes. Forest, savanna and crop were mapped best, with F1 scores up to 85.0% when combining multi-modal data, compared to 81.6% reached by DeepLab. Furthermore, in a qualitative analysis, we highlight that the classifiers sometimes outperform the inaccurate labels.

Time-series land cover mapping and urban expansion analysis using OpenStreetMap data and remote sensing big data: A case study of Guangdong-Hong Kong-Macao Greater Bay Area, China

• Optimized training samples from OpenStreetMap facilitate large-scale land cover mapping. • Sample optimization and transfer, and consistency modification strategies improve mapping results. • The proposed workflow achieves satisfactory performance and is transferable to other regions. Accurate land cover mapping provides important scientific support for ecological environment protection and sustainable urban development. However, given the high expense of acquiring training samples and the difficulty of fully utilizing remote sensing big data, large-scale time-series land cover mapping remains to be a challenge. To address the issue, we proposed a novel time-series large-scale mapping approach that obtains high-quality training samples from OpenStreetMap volunteered data and transfers them for land cover mapping in historical years. Relying on the data archived on the Google Earth Engine platform, we constructed a discriminating feature set that contains spectrum, texture, and backscatter coefficient, among others. Taking the Guangdong-Hong Kong-Macao Greater Bay Area as a case area, the annual land cover maps after spatio-temporal consistency modification from 1986 to 2021 were obtained. The validation samples proved that the derived land cover classification results have high accuracy. We verified the reliability and superiority of the proposed approach by comparing our results with existing land cover products. The proposed mapping approach owns great transferability, and the mapping results provide a valid decision-making basis for urban planning.

Mapping Land Cover Types for Highland Andean Ecosystems in Peru Using Google Earth Engine

Highland Andean ecosystems sustain high levels of floral and faunal biodiversity in areas with diverse topography and provide varied ecosystem services, including the supply of water to cities and downstream agricultural valleys. Google (™) has developed a product specifically designed for mapping purposes (Earth Engine), which enables users to harness the computing power of a cloud-based solution in near-real time for land cover change mapping and monitoring. We explore the feasibility of using this platform for mapping land cover types in topographically complex terrain with highly mixed vegetation types (Nor Yauyos Cochas Landscape Reserve located in the central Andes of Peru) using classification machine learning (ML) algorithms in combination with different sets of remote sensing data. The algorithms were trained using 3601 sampling pixels of (a) normalized spectral bands between the visible and near infrared spectrum of the Landsat 8 OLI sensor for the 2018 period, (b) spectral indices of vegetation, soil, water, snow, burned areas and bare ground and (c) topographic-derived indices (elevation, slope and aspect). Six ML algorithms were tested, including CART, random forest, gradient tree boosting, minimum distance, naïve Bayes and support vector machine. The results reveal that ML algorithms produce accurate classifications when spectral bands are used in conjunction with topographic indices, resulting in better discrimination among classes with similar spectral signatures such as pajonal (tussock grass-dominated cover) and short grasses or rocky groups, and moraines, agricultural and forested areas. The model with the highest explanatory power was obtained from the combination of spectral bands and topographic indices using the random forest algorithm (Kappa = 0.81). Our study presents a first approach of its kind in topographically complex Cordilleran terrain and we show that GEE is particularly useful in large-scale land cover mapping and monitoring in mountainous ecosystems subject to rapid changes and conversions, with replicability and scalability to other areas with similar characteristics.

Hidden Markov Models for Annual Land Cover Mapping—Increasing Temporal Consistency and Completeness

This article aims at investigating the hidden Markov model (HMM) approach for the automated processing of classified satellite images for land cover and land-use change (LCLUC). HMM’s account for transitions between classes at the same location, but that cannot be directly observed due to classification errors. Using a set of transition and emission probabilities, HMM’s allow filtering out errors and recovering the actual sequence of LCLUC, which are typically overestimated when directly estimated from the classified images. After presenting the HMM framework, the methodology is illustrated on three 300-m annual time series of classified images from 2003 to 2019 over <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$756\times756$ </tex-math></inline-formula> km <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sup> areas in Brazil, People’s Republic of China, and Mali. It is shown how the emission and transition probabilities can be estimated from these time series using a simple Viterbi training, alleviating computationally demanding algorithms. Special attention is paid to the processing of missing observations caused by clouds. Combining these three datasets with a simulation study, it is concluded that the HMM emission and transition probabilities can be estimated with low biases and variances thanks to the vast number (hundreds of thousands) of pixels at hand. The speed of the Viterbi training and decoding steps makes it possible to consider large-scale land cover mapping at moderate or even high spatial resolution as long as the legend of the LCLUC involves a reasonable number of classes like the six main Intergovernmental Panel on Climate Change (IPCC) land categories.

A Scalable High-Performance Unsupervised System for Producing Large-Scale HR Land Cover Maps: The Italian Country Case Study

This paper presents an operational system for the automatic production of HR large-scale Land Cover (LC) maps in a fast, efficient and unsupervised manner. This is based on a scalable and parallelizable tile-based approach, which does not require the collection of new training data. The method leverages the complementary information provided by existing LC maps and recent acquisitions of HR Earth Observation (EO) images to: (i) identify map units that have the highest probability of being correctly associated with their labels, and (ii) exploit the obtained “weak” training set to produce an updated HR LC map by classifying the recently acquired EO data. Both steps, performed at tile level, can be implemented on a High Performance Computing (HPC) environment, which simultaneously process all required tiles (independently of each other) for the entire study area. The method was tested considering the publicly available 2018 Corine Land Cover (CLC) Map having a minimum mapping unit of 25Ha and the Sentinel-2 images to generate a HR LC map of Italy. The obtained map has a spatial resolution of 10 m and presents the nine major LC types (i.e., “Artificial Land,” “Bareland,” “Grassland,” “Cropland,” “Broadleaves,” “Conifers,” “Snow,” “Water” and “Shrubland”). Validation was performed using the 2018 European Land Use and Coverage Area Frame Survey (LUCAS) database made up of <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">in-situ</i> data. The Overall Accuracy (OA) achieved for the Northern, Southern and Central part of Italy and Italian Islands are 91.29%, 91.63%, 92.21% and 91.06%, respectively.

Transferable Deep Learning from Time Series of Landsat Data for National Land-Cover Mapping with Noisy Labels: A Case Study of China

Large-scale land-cover classification using a supervised algorithm is a challenging task. Enormous efforts have been made to manually process and check the production of national land-cover maps. This has led to complex pre- and post-processing and even the production of inaccurate mapping products from large-scale remote sensing images. Inspired by the recent success of deep learning techniques, in this study we provided a feasible automatic solution for improving the quality of national land-cover maps. However, the application of deep learning to national land-cover mapping remains limited because only small-scale noisy labels are available. To this end, a mutual transfer network MTNet was developed. MTNet is capable of learning better feature representations by mutually transferring pre-trained models from time-series of data and fine-tuning current data. An interactive training strategy such as this can effectively alleviate the effects of inaccurate or noisy labels and unbalanced sample distributions, thus yielding a relatively stable classification system. Extensive experiments were conducted by focusing on several representative regions to evaluate the classification results of our proposed method. Quantitative results showed that the proposed MTNet outperformed its baseline model about 1%, and the accuracy can be improved up to 6.45% compared with the model trained by the training set of another year. We also visualized the national classification maps generated by MTNet for two different time periods to quantitatively analyze the performance gain. It was concluded that the proposed MTNet provides an efficient method for large-scale land cover mapping.

Spatial Resolution Enhancement for Large-Scale Land Cover Mapping via Weakly Supervised Deep Learning

Multispectral satellite imagery is the primary data source for monitoring land cover change and characterizing land cover globally. However, the consistency of land cover monitoring is limited by the spatial and temporal resolutions of the acquired satellite images. The public availability of daily high-resolution images is still scarce. This paper aims to fill this gap by proposing a novel spatiotemporal fusion method to enhance daily low spatial resolution land cover mapping using a weakly supervised deep convolutional neural network. We merge Sentinel images and moderate resolution imaging spectroradiometer (MODIS )-derived thematic land cover maps under the application background of massive remote sensing data and the large spatial resolution gaps between MODIS data and Sentinel images. The neural network training was conducted on the public data set SEN12MS, while the validation and testing used ground truth data from the 2020 IEEE Geoscience and Remote Sensing Society data fusion contest. The proposed data fusion method shows that the synthesized land cover map has significantly higher spatial resolution than the corresponding MODIS-derived land cover map. The ensemble approach can be implemented for generating high-resolution time series of satellite images by fusing fine images from Sentinel-1 and -2 and daily coarse images from MODIS.

Optimal and robust vegetation mapping in complex environments using multiple satellite imagery: Application to mangroves in Southeast Asia

A band selection model was described for efficient and accurate remotely-sensed vegetation mapping in cloudy mixed-vegetation areas, demonstrated with an application on mapping mangroves in Southeast Asia (SE Asia). We show how to use multi-source satellite imagery and Cloud Computing Platforms to improve mapping and computational efficiency in complex environments. A key element of the method relies upon field surveys to establish a detailed sample database that includes easily-confused land cover. The Maximal Separability and Information (MSI) model was developed to select key bands for target land cover classification from multiple satellite imagery based on two principles: 1. maximize separability of the target cover from other land cover; and 2. maximize and prioritize information from band combinations. Application of the MSI model to map mangroves in SE Asia using three optical and SAR data systems (Landsat OLI, Sentinel-2 and Sentinel-1) showed: 1. Sentinel-2 is better at classifying mangrove than Landsat and Sentinel-1; and 2. SWIR, NIR and Red bands (with SWIR in particular) are effective in separating mangrove from other vegetation. The MSI-mapped mangroves showed lower computation cost compared to using all bands from individual satellites, and higher accuracy (above 90%) when applied to SE Asia. It was robust in tolerating smaller sample sizes, thereby demonstrating computational feasibility and substantial improvements with the MSI model for large-scale land cover mapping in complex environments.

WEAKLY SUPERVISED SEMANTIC SEGMENTATION OF SATELLITE IMAGES FOR LAND COVER MAPPING – CHALLENGES AND OPPORTUNITIES

Abstract. Fully automatic large-scale land cover mapping belongs to the core challenges addressed by the remote sensing community. Usually, the basis of this task is formed by (supervised) machine learning models. However, in spite of recent growth in the availability of satellite observations, accurate training data remains comparably scarce. On the other hand, numerous global land cover products exist and can be accessed often free-of-charge. Unfortunately, these maps are typically of a much lower resolution than modern day satellite imagery. Besides, they always come with a significant amount of noise, as they cannot be considered ground truth, but are products of previous (semi-)automatic prediction tasks. Therefore, this paper seeks to make a case for the application of weakly supervised learning strategies to get the most out of available data sources and achieve progress in high-resolution large-scale land cover mapping. Challenges and opportunities are discussed based on the SEN12MS dataset, for which also some baseline results are shown. These baselines indicate that there is still a lot of potential for dedicated approaches designed to deal with remote sensing-specific forms of weak supervision.

Comparing Fully Deep Convolutional Neural Networks for Land Cover Classification with High-Spatial-Resolution Gaofen-2 Images

Land cover is an important variable of the terrestrial ecosystem that provides information for natural resources management, urban sprawl detection, and environment research. To classify land cover with high-spatial-resolution multispectral remote sensing imagery is a difficult problem due to heterogeneous spectral values of the same object on the ground. Fully convolutional networks (FCNs) are a state-of-the-art method that has been increasingly used in image segmentation and classification. However, a systematic quantitative comparison of FCNs on high-spatial-multispectral remote imagery was not yet performed. In this paper, we adopted the three FCNs (FCN-8s, Segnet, and Unet) for Gaofen-2 (GF2) satellite imagery classification. Two scenes of GF2 with a total of 3329 polygon samples were used in the study area and a systematic quantitative comparison of FCNs was conducted with red, green, blue (RGB) and RGB+near infrared (NIR) inputs for GF2 satellite imagery. The results showed that: (1) The FCN methods perform well in land cover classification with GF2 imagery, and yet, different FCNs architectures exhibited different results in mapping accuracy. The FCN-8s model performed best among the Segnet and Unet architectures due to the multiscale feature channels in the upsampling stage. Averaged across the models, the overall accuracy (OA) and Kappa coefficient (Kappa) were 5% and 0.06 higher, respectively, in FCN-8s when compared with the other two models. (2) High-spatial-resolution remote sensing imagery with RGB+NIR bands performed better than RGB input at mapping land cover, and yet the advantage was limited; the OA and Kappa only increased an average of 0.4% and 0.01 in the RGB+NIR bands. (3) The GF2 imagery provided an encouraging result in estimating land cover based on the FCN-8s method, which can be exploited for large-scale land cover mapping in the future.

A global cloud free pixel- based image composite from Sentinel-2 data.

Large-scale land cover classification from satellite imagery is still a challenge due to the big volume of data to be processed, to persistent cloud-cover in cloud-prone areas as well as seasonal artefacts that affect spatial homogeneity. Sentinel-2 times series from Copernicus Earth Observation program offer a great potential for fine scale land cover mapping thanks to high spatial and temporal resolutions, with a decametric resolution and five-day repeat time. However, the selection of best available scenes, their download together with the requirements in terms of storage and computing resources pose restrictions for large-scale land cover mapping. The dataset presented in this paper corresponds to global cloud-free pixel based composite created from the Sentinel-2 data archive (Level L1C) available in Google Earth Engine for the period January 2017- December 2018. The methodology used for generating the image composite is described and the metadata associated with the 10 m resolution dataset is presented. The data with a total volume of 15 TB is stored on the Big Data platform of the Joint Research Centre. It can be downloaded per UTM grid zone, loaded into GIS clients and displayed easily thanks to pre-computed overviews.

2020 IEEE GRSS Data Fusion Contest: Global Land Cover Mapping With Weak Supervision [Technical Committees

The 2020 Data Fusion Contest, organized by the IEEE Geoscience and Remote Sensing Society (GRSS) Image Analysis and Data Fusion Technical Committee (IADF TC) and the Technical University of Munich, aims to promote research in automatic large-scale land cover mapping from globally available multimodal satellite data with weak supervision.

Producing consistent visually interpreted land cover reference data: learning from feedback

ABSTRACT Reference data for large-scale land cover map are commonly acquired by visual interpretation of remotely sensed data. To assure consistency, multiple images are used, interpreters are trained, sites are interpreted by several individuals, or the procedure includes a review. But little is known about important factors influencing the quality of visually interpreted data. We assessed the effect of multiple variables on land cover class agreement between interpreters and reviewers. Our analyses concerned data collected for validation of a global land cover map within the Copernicus Global Land Service project. Four cycles of visual interpretation were conducted, each was followed by review and feedback. Each interpreted site element was labelled according to dominant land cover type. We assessed relationships between the number of interpretation updates following feedback and the variables grouped in personal, training, and environmental categories. Variable importance was assessed using random forest regression. Personal variable interpreter identifier and training variable timestamp were found the strongest predictors of update counts, while the environmental variables complexity and image availability had least impact. Feedback loops reduced updating and hence improved consistency of the interpretations. Implementing feedback loops into the visually interpreted data collection increases the consistency of acquired land cover reference data.

Fusing Multiseasonal Sentinel-2 Imagery for Urban Land Cover Classification With Multibranch Residual Convolutional Neural Networks

Exploiting multitemporal Sentinel-2 images for urban land cover classification has become an important research topic, since these images have become globally available at relatively fine temporal resolution, thus offering great potential for large-scale land cover mapping. However, appropriate exploitation of the images needs to address problems such as cloud cover inherent to optical satellite imagery. To this end, we propose a simple yet effective decision-level fusion approach for urban land cover prediction from multiseasonal Sentinel-2 images, using the state-of-the-art residual convolutional neural networks (ResNet). We extensively tested the approach in a cross-validation manner over a seven-city study area in central Europe. Both quantitative and qualitative results demonstrated the superior performance of the proposed fusion approach over several baseline approaches, including observation- and feature-level fusion.

IRANIAN LAND COVER MAPPING USING LANDSAT-8 IMAGERY AND RANDOM FOREST ALGORITHM

Abstract. Land cover classification is important for various environmental assessments. The opportunity of imaging the Earth’s surface makes remote sensing techniques efficient approaches for land cover classification. The only country-wide land cover map of Iran was produced by the Iranian Space Agency (ISA) using low spatial resolution Moderate Resolution Imaging Spectroradiometer (MODIS) imagery and a basic classification method. Thus, it is necessary to produce a more accurate map using advanced remote sensing and machine learning techniques. In this study, multi-temporal Landsat-8 data (1,321 images) were inserted into a Random Forest (RF) algorithm to classify the land cover of the entire country into 13 categories. To this end, all steps, including pre-processing, classification, and accuracy assessment were implemented in the Google Earth Engine (GEE) platform. The overall classification accuracy and Kappa Coefficient obtained from the Iran-wide map were 74% and 0.71, respectively, indicating the high potential of the proposed method for large-scale land cover mapping.

Large-scale probabilistic identification of boreal peatlands using Google Earth Engine, open-access satellite data, and machine learning.

Freely-available satellite data streams and the ability to process these data on cloud-computing platforms such as Google Earth Engine have made frequent, large-scale landcover mapping at high resolution a real possibility. In this paper we apply these technologies, along with machine learning, to the mapping of peatlands–a landcover class that is critical for preserving biodiversity, helping to address climate change impacts, and providing ecosystem services, e.g., carbon storage–in the Boreal Forest Natural Region of Alberta, Canada. We outline a data-driven, scientific framework that: compiles large amounts of Earth observation data sets (radar, optical, and LiDAR); examines the extracted variables for suitability in peatland modelling; optimizes model parameterization; and finally, predicts peatland occurrence across a large boreal area (397, 958 km2) of Alberta at 10 m spatial resolution (equalling 3.9 billion pixels across Alberta). The resulting peatland occurrence model shows an accuracy of 87% and a kappa statistic of 0.57 when compared to our validation data set. Differentiating peatlands from mineral wetlands achieved an accuracy of 69% and kappa statistic of 0.37. This data-driven approach is applicable at large geopolitical scales (e.g., provincial, national) for wetland and landcover inventories that support long-term, responsible resource management.