Resolution Mapping Research Articles

WaveletFormerNet: A Transformer-based wavelet network for real-world non-homogeneous and dense fog removal

Although deep convolutional neural networks have achieved remarkable success in removing synthetic fog, it is essential to be able to process images taken in complex foggy conditions, such as dense or non-homogeneous fog, in the real world. However, the haze distribution in the real world is complex, and downsampling can lead to color distortion or loss of detail in the output results as the resolution of a feature map or image resolution decreases. Moreover, the over-stacking of convolutional blocks might increase the model complexity. In addition to the challenges of obtaining sufficient training data, overfitting can also arise in deep learning techniques for foggy image processing, which can limit the generalization abilities of the model, posing challenges for its practical applications in real-world scenarios. Considering these issues, this paper proposes a Transformer-based wavelet network (WaveletFormerNet) for real-world foggy image recovery. We embed the discrete wavelet transform into the Vision Transformer by proposing the WaveletFormer and IWaveletFormer blocks, aiming to alleviate texture detail loss and color distortion in the image due to downsampling. We introduce parallel convolution in the Transformer block, which allows for the capture of multi-frequency information in a lightweight mechanism. Such a structure reduces computational expenses and improves the effectiveness of the network. Additionally, we have implemented a feature aggregation module (FAM) to maintain image resolution and enhance the feature extraction capacity of our model, further contributing to its impressive performance in real-world foggy image recovery tasks. Through extensive experiments on real-world fog datasets, we have demonstrated that our WaveletFormerNet achieves superior performance compared to state-of-the-art methods, as shown through quantitative and qualitative evaluations of minor model complexity. Additionally, our satisfactory results on real-world dust removal and application tests showcase the superior generalization ability and improved performance of WaveletFormerNet in computer vision-related applications compared to existing state-of-the-art methods, further confirming our proposed approach's effectiveness and robustness. Our code is available at https://github.com/shengli666666/WaveletFormerNet.

A modular framework for FAIR shallow landslide susceptibility mapping based on machine learning

Machine learning has grown in popularity in the past few years for susceptibility and hazard mapping tasks. Necessary steps for the generation of a susceptibility or hazard map are repeatedly implemented in new studies. We present a Random Forest classifier-based landslide susceptibility and hazard mapping framework to facilitate future mapping studies using machine learning. The framework, as a piece of software, follows the FAIR paradigm, and hence is set up as a transparent, reproducible and modularly extensible workflow. It contains pre-implemented steps from conceptualisation to map generation, such as the generation of input datasets. The framework can be applied to different areas of interest using different environmental features and is also flexible in terms of the desired scale and resolution of the final map. To demonstrate the functionality and validity of the framework, and to explore the challenges and limitations of Random Forest-based susceptibility and hazard mapping, we apply the framework to a test case. This test case conveys the influence of the training dataset on the generated susceptibility maps in terms of feature combination, influence of non-landslide instances and representativeness of the training data with respect to the area of interest. A comparison of the test case results with the literature shows that the framework works reliably. Furthermore, the results obtained in this study complement the findings of previous studies that demonstrate the sensitivity of the training process to the training data, particularly in terms of its representativeness.

Improving pluvial flood mapping resolution of large coarse models with deep learning

ABSTRACT Deep learning (DL) models are a promising complement to hydrodynamic models. However, the application of DL for detailed predictions in large domains has not yet been tested. We aim to narrow this gap by improving flood mapping resolution derived from large coarse flood models. We used cGAN-Flood, a conditional generative adversarial-based model (cGAN), that showed satisfactory generalization. We demonstrate the applicability of cGAN-Flood by coupling it with mesh- and raster-based coarse models. A Hydrologic Engineering Center (HEC) River Analysis System (RAS) model (cell size of 32 m), which is mesh-based, was created for a 350 km2 watershed. In contrast, a HydroPol2D model, raster-based, was created for a 150 km2 watershed with a 15 m pixel size. We evaluated our method’s performance against a 3 m resolution HEC-RAS model in seven catchments across the cities of San Antonio and Sao Paulo. Results indicate the cGAN-Flood improved flood map accuracy, illustrating how DL can enhance flood mapping resolution.

Fusing multimodal data of nature-economy-society for large-scale urban building height estimation

The building height holds significant importance for comprehensively understanding urban morphology, enhancing urban planning, and fostering sustainable development. Although many methods using optical and SAR images have been presented for building height estimation, these methods fall short in capturing the influences of economic and social attributes on building height. In this study, we introduced a Nature-Economy-Society (NES) feature model to comprehensively represent building height information, and established a multi-scale One-Dimensional (1-D) Convolutional Neural Network for predicting building heights, referred to as NES-CNN. First, we derived the natural attributes of urban buildings from time-series Sentinel-1 SAR images and Sentinel-2 multispectral images, as well as World Settlement Footprint (WSF) data and Digital Elevation Model (DEM), economic attributes from nighttime light and Gross Domestic Product (GDP) data, and social function attributes from Points of Interest (POI) data. Second, an autoencoder is employed to reduce the dimensionality of the high-dimensional natural attribute features, minimizing data redundancy. Finally, the multi-scale 1-D CNN model is presented to explore the correlations between the multi-source and heterogeneous NES features and building height information, facilitating the prediction of building height. In experiments, we applied the proposed method to estimate building heights in Beijing and Shanghai at a spatial resolution of 10 m. The results indicated that for Beijing, the RMSE, MAE, and R values are 6.93 m, 4.41 m, and 0.84, respectively, while for Shanghai, these values are 7.57 m, 5.38 m, and 0.80, respectively. The addition of social and economic attribute information decreases the RMSE by 6 % in both Beijing and Shanghai compared with using only natural attributes. In comparison to existing studies at the same mapping resolution, RMSE decreases by 39 % for Beijing and 51 % for Shanghai. The innovative and inspiring nature of this study lies in its application to large-scale building height estimation.

Spatially Resolved High Voltage Kelvin Probe Force Microscopy: A Novel Avenue for Examining Electrical Phenomena at Nanoscale

AbstractKelvin probe force microscopy (KPFM) is a well‐established scanning probe technique, used to measure surface potential accurately; it has found extensive use in the study of a range of materials phenomena. In its conventional form, KPFM frustratingly precludes imaging samples or scenarios where large surface potential or surface potential gradients exist outside the typical ±10 V window. If the potential regime measurable via KPFM can be expanded, to enable precise and reliable metrology, through a high voltage KPFM (HV‐KPFM) adaptation, it can open up pathways toward a range of novel experiments, where the detection limit of regular KPFM has so far prevented the use of the technique. In this work, HV‐KPFM is realized and shown to be capable of measuring large surface potential and potential gradients with accuracy and precision. The technique is employed to study a range of materials (positive temperature coefficient of resistivity ceramics, charge storage fluoropolymers, and pyroelectrics) where accurate, spatially resolved mapping of surface potential within high voltage regime facilitates novel physical insight. The results demonstrate that HV‐KPFM can be used as an effective tool to fill in existing gaps in surface potential measurements while also opening routes for novel studies in materials physics.

Cryo-EM Map Anisotropy Can Be Attenuated by Map Post-Processing and a New Method for Its Estimation.

One of the most important challenges in cryogenic electron microscopy (cryo-EM) is the substantial number of samples that exhibit preferred orientations, which leads to an uneven coverage of the projection sphere. As a result, the overall quality of the reconstructed maps can be severely affected, as manifested by the presence of anisotropy in the map resolution. Several methods have been proposed to measure the directional resolution of maps in tandem with experimental protocols to address the problem of preferential orientations in cryo-EM. Following these works, in this manuscript we identified one potential limitation that may affect most of the existing methods and we proposed an alternative approach to evaluate the presence of preferential orientations in cryo-EM reconstructions. In addition, we also showed that some of the most recently proposed cryo-EM map post-processing algorithms can attenuate map anisotropy, thus offering alternative visualization opportunities for cases affected by moderate levels of preferential orientations.

Global and single-nucleotide resolution detection of 7-methylguanosine in RNA

ABSTRACT RNA modifications, including N-7-methylguanosine (m7G), are pivotal in governing RNA stability and gene expression regulation. The accurate detection of internal m7G modifications is of paramount significance, given recent associations between altered m7G deposition and elevated expression of the methyltransferase METTL1 in various human cancers. The development of robust m7G detection techniques has posed a significant challenge in the field of epitranscriptomics. In this study, we introduce two methodologies for the global and accurate identification of m7G modifications in human RNA. We introduce borohydride reduction sequencing (Bo-Seq), which provides base resolution mapping of m7G modifications. Bo-Seq achieves exceptional performance through the optimization of RNA depurination and scission, involving the strategic use of high concentrations of NaBH4, neutral pH and the addition of 7-methylguanosine monophosphate (m7GMP) during the reducing reaction. Notably, compared to NaBH4-based methods, Bo-Seq enhances the m7G detection performance, and simplifies the detection process, eliminating the necessity for intricate chemical steps and reducing the protocol duration. In addition, we present an antibody-based approach, which enables the assessment of m7G relative levels across RNA molecules and biological samples, however it should be used with caution due to limitations associated with variations in antibody quality between batches. In summary, our novel approaches address the pressing need for reliable and accessible methods to detect RNA m7G methylation in human cells. These advancements hold the potential to catalyse future investigations in the critical field of epitranscriptomics, shedding light on the complex regulatory roles of m7G in gene expression and its implications in cancer biology.

The digital climate atlas of the Canary Islands: A tool to improve knowledge of climate and temperature and precipitation trends in the Atlantic islands

This paper presents the Interactive Digital Climate Atlas of the Canary Islands, a web portal based on geographic information systems for the dissemination of climate information. It provides citizens and public and private administrations with high spatial resolution maps (100 m) of different climate variables such as temperature, rainfall, relative humidity, cloudiness frequency and wind velocity. Also available are Köppen climate classification maps and decadal trends of temperature and precipitation in a context of climate change. This spatial resolution is suitable for the detailed climatic description of small and orographically complex islands. For this purpose, climate series from databases of different public institutions as well as data from remote sensing and reanalysis systems were employed. A statistical process involving data filtering and the detection and correction of inhomogeneous segments was carried out as a preparatory step. Then, a spatial interpolation model was developed using a multiple linear regression method for the generation of the high spatial resolution climate cartography. The results are annual and monthly maps of the climate variables in Canary Islands at 100 m spatial resolution and with different time periods depending on the availability of data on the different variables, but in most cases between 1991 and 2020. Following the European Directive INSPIRE (Directive 2007/2/CE, Infrastructure for Spatial Information in Europe), both map visualisation and downloading in GeoTIFF and NetCDF format are permitted, as well as consultation of the data and metadata of the climatic series used.

Enhance Road Detection Data Processing of LiDAR Point Clouds to Specifically Identify Unmarked Gravel Rural Roads

Abstract Gravel roads lack standardized features such as curbs or painted lines, presenting detection challenges to autonomous vehicles. Global positioning service (GPS) and high resolution maps may not be reliable for navigation of gravel roads, as some roads may only be width of the vehicle and GPS may not be accurate enough. Normal distribution transform (NDT) LiDAR scan matching may be insufficient for navigating on gravel roads as there may not be enough geometrically distinct features for reliable scan matching. This paper examined a method of classifying scanning LiDAR spatial and remission data features for explicit detection of unmarked gravel road surfaces. Exploration of terrain classification using high resolution scanning LiDAR data of specific road surfaces may allow for predicting gravel road boundary locations potentially enabling confident autonomous operations on gravel roads. The principal outcome of this work was a method for gravel road terrain detection using LiDAR data for the purpose of predicting potential road boundary locations. Random decision forests were trained using scanning LiDAR data terrain classification to detect unmarked gravel and asphalt surfaces. It was found that a true-positive accuracy for gravel and asphalt surfaces was 75% and 87%, respectively, at an estimated rate of 13 ms per 360 deg scan. Overlapping results between manually projected and actual road surface areas resulted in 93% intersecting gravel road detection accuracy. Automated post-process examination of classification results yielded an true-positive gravel road detection rate of 72%.

Quantifying current and potential future impacts of balsam woolly adelgid infestation on forest biomass

Balsam woolly adelgid (Adelges piceae; BWA) is an invasive forest pest in the US whose infestation in fir forests can cause widespread tree mortality. A growing body of evidence suggests that the severity of BWA’s effects is linked to climatic conditions, where sites featuring seasonally warmer temperatures tend to demonstrate higher degrees of insect-induced forest degradation. Thus, a warming climate may promote the expansion of BWA-driven damage, particularly in highly susceptible subalpine fir (Abies lasiocarpa) forests. The degree to which climate change may foster BWA infestation and the relationship between current and future infestation damage severity are largely unknown. To understand BWA’s potential climate-driven impacts on subalpine fir forests now and in the future, we built a field-validated, spatially explicit predictive model that estimates BWA infestation severity as a function of locally downscaled temperature variables. Using only four seasonal temperature variables, our model was able to explain over 78% of variance in measured severity. We applied that model to the prediction of severity for five different time periods (2020, 2040, 2060, 2080, and 2100), with the 2040 and onward predictions being driven by four separate climatic shared socioeconomic pathways (SSP126, SSP245, SSP370, and SSP585). These high resolution (∼30 m) maps were compared to spatially coincident modeled estimates of subalpine fir aboveground biomass, the results of which provide valuable insight into the degree to which BWA may drive forest degradation, loss of ecosystem services, and increase in available fuels. In northern Utah, the southeasternmost extent of BWA infestation in the western US, the two major mountain ranges, Wasatch and Uinta, demonstrated very different climatic susceptibility to infestation damage, with the former being highly exposed to BWA-induced degradation under even moderate climate projections, and the latter being fairly robust against BWA’s effects in all but the most extreme climate scenarios. In 2020, 41% of the study area’s subalpine fir biomass is climatically exposed to some level of damage from BWA. By 2100, even under moderate climate projections (SSP245), 79% will be exposed, with 37% predicted to feature relatively high severity. By placing our results within the context of the Resist, Accept, Direct land management framework, we offer spatially explicit guidance to inform current and future silvicultural practices to mitigate future damage to subalpine fir forests at the leading edge of BWA expansion in the western US.

High resolution maps of chromatin reorganization through mouse meiosis reveal novel features of the 3D meiotic structure.

When germ cells transition from the mitotic cycle into meiotic prophase I (MPI), chromosomes condense into an array of chromatin loops that are required to promote homolog pairing and genetic recombination. To identify the changes in chromosomal conformation, we isolated nuclei on a trajectory from spermatogonia to the end of MPI. At each stage along this trajectory, we built genomic interaction maps with the highest temporal and spatial resolution to date. The changes in chromatin folding coincided with a concurrent decline in mitotic cohesion and a rise in meiotic cohesin complexes. We found that the stereotypical large-scale A and B compartmentalization was lost during meiotic prophase I alongside the loss of topological associating domains (TADs). Still, local subcompartments were detected and maintained throughout meiosis. The enhanced Micro-C resolution revealed that, despite the loss of TADs, higher frequency contact sites between two loci were detectable during meiotic prophase I coinciding with CTCF bound sites. The pattern of interactions around these CTCF sites with their neighboring loci showed that CTCF sites were often anchoring the meiotic loops. Additionally, the localization of CTCF to the meiotic axes indicated that these anchors were at the base of loops. Strikingly, even in the face of the dramatic reconfiguration of interphase chromatin into a condensed loop-array, the interactions between regulatory elements remained well preserved. This establishes a potential mechanism for how the meiotic chromatin maintains active transcription within a highly structured genome. In summary, the high temporal and spatial resolution of these data revealed previously unappreciated aspects of mammalian meiotic chromatin organization.

Genetic control of DNA methylation is largely shared across European and East Asian populations

DNA methylation is an ideal trait to study the extent of the shared genetic control across ancestries, effectively providing hundreds of thousands of model molecular traits with large QTL effect sizes. We investigate cis DNAm QTLs in three European (n = 3701) and two East Asian (n = 2099) cohorts to quantify the similarities and differences in the genetic architecture across populations. We observe 80,394 associated mQTLs (62.2% of DNAm probes with significant mQTL) to be significant in both ancestries, while 28,925 mQTLs (22.4%) are identified in only a single ancestry. mQTL effect sizes are highly conserved across populations, with differences in mQTL discovery likely due to differences in allele frequency of associated variants and differing linkage disequilibrium between causal variants and assayed SNPs. This study highlights the overall similarity of genetic control across ancestries and the value of ancestral diversity in increasing the power to detect associations and enhancing fine mapping resolution.

Abstract 4900: Spatial Atlas of Human Anatomy (SAHA): A subcellular, multiscale spatial odyssey of immune and gastrointestinal tissues and tumors

Abstract The Spatial Atlas of Human Anatomy (SAHA) is a foundational effort to map 250 million cells and transcriptomes and proteomes of 30 non-diseased organs from healthy adults at two spatial scales: whole transcriptome of histological features (50 µm to 2 mm), and 1,000-plex RNA and 64-plex protein panels at spatial subcellular resolution (50 nm across 1 cm2). The project aims to establish and validate best practices in experimental design, sample processing, data analysis, and data standards for high-content spatial analysis across multiple human organs at whole transcriptome and proteome levels. The profiled samples will capture variability across genders and ancestries. All results, including raw and processed data, will be made available to the scientific community through the SAHA data portal and AtoMxTM Spatial Informatics Platform. Here we present the part of Phase I data collected specifically around immune (bone marrow and lymph node) and gastrointestinal tissues (ileum, appendix, colon, and liver), and compare these data to colon and liver cancer samples. On serial sections, we perform spatial whole transcriptome data (with GeoMx® Digital Spatial Profiler) and 1,000-plex RNA profiles and 64-plex protein profiles collected by the CosMx™ Spatial Molecular Imager (SMI) matched to the exact shape of functional histological organ features from H&E and immunofluorescence stainings. Our data, which spans ~1.5 million cells collected across ~1,100 fields of view, enables the highest-ever subcellular resolution maps of cell types, lineage states, metabolic capacity, cellular neighborhoods, subcellular movements of organelles, and spatially resolved (and novel) ligand-receptor interactions across normal and cancer tissues. Through comparing these data to several cancer samples collected across multiple spatial platforms, we show how spatial organ atlasing at multiple scales can uncover unique insights into organ development, health, and cancer. We also show how the SAHA data can serve as a benchmark reference for spatial precision medicine. Citation Format: Jiwoon Park, Roberto Gregorio, Erika Hissong, Sanjay Patel, Brian Robinson, Fabio Socciarelli, Evenlyn Metzger, Yan Liang, Jason Reeves, Joseph Beechem, Olivier Elemento, Alicia Alonso, Shauna Houlihan, Robert Schwartz, Christopher E. Mason. Spatial Atlas of Human Anatomy (SAHA): A subcellular, multiscale spatial odyssey of immune and gastrointestinal tissues and tumors [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2024; Part 1 (Regular Abstracts); 2024 Apr 5-10; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2024;84(6_Suppl):Abstract nr 4900.

Three decades of spatiotemporal dynamics in forest biomass density in the Qinba Mountains

The forest ecosystem plays a pivotal role in the global carbon cycle and is crucial for investigating atmospheric carbon exchanges. Forest biomass, a fundamental quantitative measure of the forest ecosystem, serves as a critical indicator for forest carbon stocks and carbon sequestration capacity. This study utilizes the GIMMS NDVI3g dataset to downscale forest inventory data spanning from 1989 to 2018, creating a 1 km resolution map of forest biomass density in the Qinba Mountains. The density initially decreased but has been increasing since 2004. The northern region of the Qinba Mountains exhibits a high forest biomass density (>100 Mg/hm2), while the southern region has relatively lower biomass density. This study provides the longest-term estimation of forest biomass density in the Qinba Mountains to date. It serves as a foundation for regional-scale forest carbon sequestration management and carbonization decision-making. This research is of significant importance for enhancing understanding of regional carbon cycling and supporting sustainable ecological development.

3D attention-focused pure convolutional target detection algorithm for insulator defect detection

To ensure the timely detection of safety hazards in transmission lines and to enhance accurately the detection of insulator defects in complicated environments, this study proposes a 3D attention-focused pure convolutional target detection algorithm (CMYOLOv7) based on you only look once v7 (YOLOv7) for defect detection on insulators in complicated environments. Firstly, to address the incapacity of focusing on the target in the backbone feature extraction networks, this study proposes a focused pure convolutional feature extraction module (ConvSimCB) to enhance the extraction and focus capability on insulator defect features. Secondly, to solve the problem of loss of key detail feature information caused by maximum pooling in spatial pyramid pooling module (SPPCSPC), this study proposes a mixed spatial pyramid pooling module (MIXPCSPC) to retain abundant image texture detail information and increase accuracy in detection of tiny insulator defects. Finally, a lightweight generic upsampling operator (CARAFE) is introduced to enhance the feature map resolution to address image distortion caused by the Nearest Neighbor Method of upsampling. This study proposed CMYOLOv7 achieves 98.37% precision, 90.59% recall, 95.68% mean average precision (MAP), and 94% F1 score, higher than YOLOv7 by 0.61%, 8.64%, 5.41%, and 5%.

High-resolution extracellular pH imaging of liver cancer with multiparametric MR using Deep Image Prior.

Noninvasive extracellular pH (pHe) mapping with Biosensor Imaging of Redundant Deviation in Shifts (BIRDS) using MR spectroscopic imaging (MRSI) has been demonstrated on 3T clinical MR scanners at 8 mm3 spatial resolution and applied to study various liver cancer treatments. Although pHe imaging at higher resolution can be achieved by extending the acquisition time, a postprocessing method to increase the resolution is preferable, to minimize the duration spent by the subject in the MR scanner. In this work, we propose to improve the spatial resolution of pHe mapping with BIRDS by incorporating anatomical information in the form of multiparametric MRI and using an unsupervised deep-learning technique, Deep Image Prior (DIP). Specifically, we used high-resolution , , and diffusion-weighted imaging (DWI) MR images of rabbits with VX2 liver tumors as inputs to a U-Net architecture to provide anatomical information. U-Net parameters were optimized to minimize the difference between the output super-resolution image and the experimentally acquired low-resolution pHe image using the mean-absolute error. In this way, the super-resolution pHe image would be consistent with both anatomical MR images and the low-resolution pHe measurement from the scanner. The method was developed based on data from 49 rabbits implanted with VX2 liver tumors. For evaluation, we also acquired high-resolution pHe images from two rabbits, which were used as ground truth. The results indicate a good match between the spatial characteristics of the super-resolution images and the high-resolution ground truth, supported by the low pixelwise absolute error.

Cropland abandonment between 1986 and 2018 across the United States: spatiotemporal patterns and current land uses

Knowing where and when croplands have been abandoned or otherwise removed from cultivation is fundamental to evaluating future uses of these areas, e.g. as sites for ecological restoration, recultivation, bioenergy production, or other uses. However, large uncertainties remain about the location and time of cropland abandonment and how this process and the availability of associated lands vary spatially and temporally across the United States. Here, we present a nationwide, 30 m resolution map of croplands abandoned throughout the period of 1986–2018 for the conterminous United States (CONUS). We mapped the location and time of abandonment from annual cropland layers we created in Google Earth Engine from 30 m resolution Landsat imagery using an automated classification method and training data from the U.S. Department of Agriculture Cropland Data Layer. Our abandonment map has overall accuracies of 0.91 and 0.65 for the location and time of abandonment, respectively. From 1986 to 2018, 12.3 (±2.87) million hectares (Mha) of croplands were abandoned across CONUS, with areas of greatest change over the Ogallala Aquifer, the southern Mississippi Alluvial Plain, the Atlantic Coast, North Dakota, northern Montana, and eastern Washington state. The average annual nationwide abandoned area across our study period was 0.51 Mha per year. Annual abandonment peaked between 1997 and 1999 at a rate of 0.63 Mha year−1, followed by a continuous decrease to 0.41 Mha year−1 in 2009–2011. Among the abandoned croplands, 53% (6.5 Mha) changed to grassland and pasture, 18.6% (2.28 Mha) to shrubland and forest, 8.4% (1.03 Mha) to wetlands, and 4.6% (0.56 Mha) to non-vegetated lands. Of the areas that we mapped as abandoned, 19.6% (2.41 Mha) were enrolled in the Conservation Reserve Program as of 2020. Our new map highlights the long-term dynamic nature of agricultural land use and its relation to various competitive pressures and land use policies in the United States.

Wise-SrNet: a novel architecture for enhancing image classification by learning spatial resolution of feature maps

Wise-SrNet: a novel architecture for enhancing image classification by learning spatial resolution of feature maps

DeepETPicker: Fast and accurate 3D particle picking for cryo-electron tomography using weakly supervised deep learning

To solve three-dimensional structures of biological macromolecules in situ, large numbers of particles often need to be picked from cryo-electron tomograms. However, adoption of automated particle-picking methods remains limited because of their technical limitations. To overcome the limitations, we develop DeepETPicker, a deep learning model for fast and accurate picking of particles from cryo-electron tomograms. Training of DeepETPicker requires only weak supervision with low numbers of simplified labels, reducing the burden of manual annotation. The simplified labels combined with the customized and lightweight model architecture of DeepETPicker and accelerated pooling enable substantial performance improvement. When tested on simulated and real tomograms, DeepETPicker outperforms the competing state-of-the-art methods by achieving the highest overall accuracy and speed, which translate into higher authenticity and coordinates accuracy of picked particles and higher resolutions of final reconstruction maps. DeepETPicker is provided in open source with a user-friendly interface to support cryo-electron tomography in situ.

Kilometre-scale simulations over Fennoscandia reveal a large loss of tundra due to climate warming

Abstract. The Fennoscandian boreal and mountain regions harbour a wide range of vegetation types, from boreal forest to high alpine tundra and barren soils. The area is facing a rise in air temperature above the global average and changes in temperature and precipitation patterns. This is expected to alter the Fennoscandian vegetation composition and change the conditions for areal land use such as forestry, tourism and reindeer husbandry. In this study we used a unique high-resolution (3 km) climate scenario with considerable warming resulting from strongly increasing carbon dioxide emissions to investigate how climate change can alter the vegetation composition, biodiversity and availability of suitable reindeer forage. Using a dynamical vegetation model, including a new implementation of potential reindeer grazing, resulted in simulated vegetation maps of unprecedented high resolution for such a long time period and spatial extent. The results were evaluated at the local scale using vegetation inventories and for the whole area against satellite-based vegetation maps. A deeper analysis of vegetation shifts related to statistics of threatened species was performed in six “hotspot” areas containing records of rare and threatened species. In this high-emission scenario, the simulations show dramatic shifts in the vegetation composition, accelerating at the end of the century. Alarmingly, the results suggest the southern mountain alpine region in Sweden will be completely covered by forests at the end of the 21st century, making preservation of many rare and threatened species impossible. In the northern alpine regions, most vegetation types will persist but shift to higher elevations with reduced areal extent, endangering vulnerable species. Simulated potential for reindeer grazing indicates latitudinal differences, with higher potential in the south in the current climate. In the future these differences will diminish, as the potentials will increase in the north, especially for the summer grazing grounds. These combined results suggest significant shifts in vegetation composition over the present century for this scenario, with large implications for nature conservation, reindeer husbandry and forestry.