Band Data Research Articles

Study on the Evolutionary Characteristics of Post-Fire Forest Recovery Using Unmanned Aerial Vehicle Imagery and Deep Learning: A Case Study of Jinyun Mountain in Chongqing, China

Forest fires pose a significant threat to forest ecosystems, with severe impacts on both the environment and human society. Understanding the post-fire recovery processes of forests is crucial for developing strategies for species diversity conservation and ecological restoration and preventing further damage. The present study proposes applying the EAswin-Mask2former model based on semantic segmentation in deep learning using visible light band data to better monitor the evolution of burn areas in forests after fires. This model is an improvement of the classical semantic segmentation model Mask2former and can better adapt to the complex environment of burned forest areas. This model employs Swin-Transformer as the backbone for feature extraction, which is particularly advantageous for processing high-resolution images. It also includes the Contextual Transformer (CoT) Block to better capture contextual information capture and incorporates the Efficient Multi-Scale Attention (EMA) Block into the Efficiently Adaptive (EA) Block to enhance the model’s ability to learn key features and long-range dependencies. The experimental results demonstrate that the EAswin-Mask2former model can achieve a mean Intersection-over-Union (mIoU) of 76.35% in segmenting complex forest burn areas across different seasons, representing improvements of 3.26 and 0.58 percentage points, respectively, over the Mask2former models using ResNet and Swin-Transformer backbones, respectively. Moreover, this method surpasses the performance of the DeepLabV3+ and Segformer models by 4.04 and 1.75 percentage points, respectively. Ultimately, the proposed model offers excellent segmentation performance for both forest and burn areas and can effectively track the evolution of burned forests when combined with unmanned aerial vehicle (UAV) remote sensing images.

Euclid preparation

Euclid will collect an enormous amount of data during the mission’s lifetime, observing billions of galaxies in the extragalactic sky. Along with traditional template-fitting methods, numerous machine learning (ML) algorithms have been presented for computing their photometric redshifts and physical parameters (PPs), requiring significantly less computing effort while producing equivalent performance measures. However, their performance is limited by the quality and amount of input information entering the model (the features), to a level where the recovery of some well-established physical relationships between parameters might not be guaranteed – for example, the star-forming main sequence (SFMS). To forecast the reliability of Euclid photo-zs and PPs calculations, we produced two mock catalogs simulating the photometry with the UNIONS ugriz and Euclid filters. We simulated the Euclid Wide Survey (EWS) and Euclid Deep Fields (EDF), alongside two auxiliary fields. We tested the performance of a template-fitting algorithm (Phosphoros) and four ML methods in recovering photo-zs, PPs (stellar masses and star formation rates), and the SFMS on the simulated Euclid fields. To mimic the Euclid processing as closely as possible, the models were trained with Phosphoros-recovered labels and tested on the simulated ground truth. For the EWS, we found that the best results are achieved with a mixed labels approach, training the models with wide survey features and labels from the Phosphoros results on deeper photometry, that is, with the best possible set of labels for a given photometry. This imposes a prior to the input features, helping the models to better discern cases in degenerate regions of feature space, that is, when galaxies have similar magnitudes and colors but different redshifts and PPs, with performance metrics even better than those found with Phosphoros. We found no more than 3% performance degradation using a COSMOS-like reference sample or removing u band data, which will not be available until after data release DR1. The best results are obtained for the EDF, with appropriate recovery of photo-z, PPs, and the SFMS.

Archaeological cognition of the Eastern mausoleum of Qin state using integrated space-ground observation tools

The Eastern Mausoleum of Qin State is a significant component of the Qin Dynasty's royal tombs, reflecting the social development level during the Warring States period (475 BC ~ 221 BC) in China. At the onset of our investigation, we mapped the site's boundaries utilizing Corona satellite imagery and employed the Normalized Difference Vegetation Index (NDVI) and Normalized Difference Moisture Index (NDMI) to identify archaeological features. Additionally, this study is the first to propose the use of thermal infrared band data from the SDGSAT-1 satellite to explore thermal archaeological traces, demonstrating their viability for archaeological site analysis. Subsequently, electromagnetic (EM) prospection was utilized to validate the presence of an ancient burial chamber passage. Landscape monitoring and analysis of the No.1 Mausoleum were performed using Corona and Google Earth images, revealing the accuracy of the Geomancy Theory of Chinese mausoleums through dynamic remote sensing of surface changes. Furthermore, our employment of space-to-ground observational modalities and resultant Digital Elevation Models (DEM) have been used to provided new insights into the application of Remote Sensing (RS) and Geomancy in archaeology, thereby emphasizing the pivotal role of site selection in heritage preservation. This research underscores the promise of synergistic space-ground observations in both exploiting the archaeological riches of cultural heritage sites and ensuring the enduring conservation of these irreplaceable patrimonial assets.

Urban nighttime aerosol optical depth retrieval methods from SDGSAT-1 glimmer data

ABSTRACT Aerosol Optical Depth (AOD) is vital for assessing air quality and climate. Nighttime lights (NTL) provide a new data source for urban AOD inversion, but current NTL data suffer from low spatial resolution, limiting detailed AOD studies in smaller urban areas and hindering high-precision AOD algorithms. This study introduces methods for retrieving AOD using Glimmer Imager (GLI) data from the newly launched SDGSAT-1. Three methods were proposed: the standard deviation method, look-up-table method, and collaborative inversion method with VIIRS Day Night Band (DNB) data. Applied to Beijing, these methods produced a 10 m resolution AOD product. The look-up-table method achieved the highest correlation (R = 0.8876) compared to the standard deviation (R = 0.8224) and collaborative inversion (R = 0.8322) methods. The look-up-table method performed best with ample data, while the collaborative inversion method proved more reliable with limited data. These findings demonstrate the effectiveness of SDGSAT-1 GLI data in accurately retrieving nighttime AOD, offering significant potential for monitoring urban pollution.

Structural Analysis Technique and Its Validation for Large Range Mirror Refocusing System for Spaceborne Deployable Telescopes

Large aperture deployable telescopes with numerous onboard instruments are popular in the space industry for astronomy. Scientific data in various wavelength bands, from 0.6µ to 28µ, is provided by these instruments and is crucial for deep space exploration. The deployment mechanisms handle the alignment and functionality at the telescope level only. At instrument level, there is always a need for a refocusing system which can cater individual need of instrument for focusing independent to other instruments nearby. In this paper, we present a structural analysis technique for such a type of large range mirror refocusing mechanism, similar to the Near Infrared Spectrograph (NIRSpec) payload of the James Webb Space Telescope (JWST). The method relies on geometric non-linearity for large deflection, in which the structure's stiffness matrix changes as a function of loading. We have designed, analyzed structurally, and tested a 4mm range slider crank based compliant mechanism. The outcomes of linear and non-linear static analysis, computed numerically, have been compared. There are noticeable significant variations in one direction between the two analyses' results. The results are validated using appropriately designed and realized test setups, like measurements of displacement, changes in focus and image position of the optical system.

Remote sensing of turbid coastal and estuarine waters with VIIRS I (375 m) and M (750 m) bands

ABSTRACT The Visible Infrared Imaging Radiometer Suite (VIIRS) is a visible, near and shortwave, to thermal infrared multispectral scanning instrument operational on three polar orbiting satellites, Suomi-NPP, JPSS-1, and JPSS-2. In the present paper, the processing of VIIRS using ACOLITE is introduced, using the Dark Spectrum Fitting (DSF) algorithm for processing of the visible to shortwave infrared bands. ACOLITE now includes support for processing both the imaging (I) and moderate (M) resolution bands at 375 m and 750 m spatial resolution, respectively. In most conditions encountered in the present study, the SWIR bands (either I or M) are automatically selected by the DSF for performing the aerosol correction. The processing is evaluated for turbid water remote sensing via autonomous hyperspectral radiometry from four sites across coastal and estuarine waters: two sites in Belgium and one each in France and Argentina. Through analysis of hundreds of matchups between the satellite and in situ measurements, a generally good performance is found for both I and M bands, especially for bands with the largest water signal, i.e. bands between 490 and 670 nm, where on average relative differences of 10–15% were found. Reflectance biases are generally less than 0.01, with a negative sign in the green and red bands and a positive sign in the blue and NIR bands. Similar matchup results are found for the I and M red and NIR bands, with a slightly higher scatter for the NIR bands. An additional comparison with OCSSW/l2gen processing of the M band data is performed for various configurations. Overall, DSF performance is better in the visible bands, whereas l2gen outputs are more closely aligned with the in situ measurements in the NIR. On average, negative biases are found for all l2gen configurations, up to −0.02 in the blue bands. Using either the SWIR1 + 2 or SWIR1 + 3 bands for the aerosol correction gives the best performance for l2gen processing. For the three VIIRS instruments separately, the average spectral differences with in situ measurements are comparable, with the most important deviation occurring at the Suomi-NPP shortest blue bands, where DSF processing gives a larger positive bias, up to nearly 0.02. For these bands, results from l2gen correspond more closely across the three instruments – although with significant negative biases for all three sensors up to −0.02 – presumably due to the use of system vicarious calibration gains in that processor. An operational network of autonomous hyperspectral instruments provides validation data for any overpassing optical imaging satellite in its commissioning or operational phase and eliminates the need for spectral interpolation or band shifting. In the case of VIIRS specifically, the hyperspectral instruments provide adequate data for the validation of the 20, 40 and 80 nm wide bands. With three operational wide-swath instruments, which provide largely interoperable data, a high frequency of observations is available, especially for study areas at higher latitudes. The novel exploitation of the I bands is now possible, thanks to the free and open source availability of ACOLITE. The advantage of the higher resolution I band data, combined with multiple VIIRS overpasses per day, is demonstrated for mapping turbidity in nearshore regions with high spatial variabilty and for detecting under-resolved floating algae.

Large aperture microphone array measurement at operating airport for sound source model creation of J-FRAIN

In 2022, authors conducted microphone array measurements at an operating airport to create sound source models for commercial aircraft based on the noise prediction framework J-FRAIN. A similar measurement was also performed by our colleagues in 2019, but it was a trial to establish the concept of the framework and was done using a microphone array in which the number of microphones was reduced to about half. In order to improve the accuracy of the sound source models, the microphone array with a 30-meter diameter consisting of 195 microphones mounted on the ground was used in 2022. Recording-based simultaneous noise measurement using sound level meters was also performed around the airport to validate prediction results in terms of time histories of 1/3 octave band sound pressure levels. This paper describes an overview of the microphone array measurements conducted in 2022 and a case study of noise prediction around the airport using sound source models created from the data of microphone array measurements conducted in 2022. Predicted results are compared with 1/3 octave band data measured using sound level meters mounted at a height of 4 m above the ground at sites of noise monitoring stations around the airport.

Analysing Thermal Land Surface Characteristics and Urban Expansion in Kenitra City, Morocco, using Remote Sensing and GIS

Abstract This research investigates how land use and cover the affect land surface temperature in Kenitra. It employs Geographic Information System techniques and remote sensing methods to monitor and control surface temperature changes. The present study integrates remote sensing and geographic information system to process satellite images. Furthermore, it focuses, also, on analyzing land surface temperature and its correlation with land cover changes over three time periods (1989, 2000, 2020) using landsat satellite data. Thus, supervised classification is employed to map these changes and derive urban heat islands from thermal band data. This study provides a comprehensive evaluation of the relationship between land use and land surface temperature. It utilizese a landscape dynamics assessment based on multi-source and multi-sensor remote sensing technologies. Specifically, the study uses landsat satellite data (TM for 1989 and OLI-8 for 2002 and 2020) to assess the effects of land use and land cover changes on land surface temperature distribution in the Kenitra city. Spatial and statistical analyses are performed by comparing maps from remotely sensed data generated using geographic information system. Therefore, these analyses reveal various changes in the kenitra region between 1989 and 2020. The primary changes observed include an increase in built-up areas and bare ground with a decrease alongside in natural areas (vegetation, etc.). The average temperatures in 1989, 2002, and 2020 were 29°C, 31.05°C, and 32°C, respectively. The obtained results are valuable for planning and managing climate scientists, land-use planners, and researchers focusing on sustainable urbanization areas. The study’s findings will assist urban planners and policymakers in adopting appropriate measures for sustainable planning in the city to mitigate the adverse effects of land surface temperature.

Remote sensing-based green and blue agricultural water footprint estimation at the river basin scale

Since the development of Water Footprint environmental indicator, significant research on blue and green crop water use and the respective water footprint estimations has been published. Such research is commonly approached using different methodologies that leverage tabulated values for crop development characterisation, while studies based on remote sensing data are less abundant, despite crop monitoring using remote sensing-based vegetation indices having demonstrated great capabilities and operability. To help fill this gap, we present a methodology that uses a remote sensing vegetation index time series from Sentinel-2 satellite near infra-red and red spectral bands data to derive basal crop coefficient time series to subsequently be used under the Remote Sensing-based Soil Water Balance approach that follows the globally operative FAO56 procedure. It provides pixel-based temporal and spatially distributed estimations of net irrigation requirements and adjusted crop evapotranspiration, with the aim being to divide up the latter and estimate the remote sensing-based green and blue crop water use and the subsequent green and blue water footprint. This is all done under the Agricultural Water Footprint Assessment framework for a growing crop or tree. This methodology was applied over a large, crop-diverse Spanish river basin district (Júcar) and across two different climatological years (humid vs. dry). Its feasibility was demonstrated by the acceptable behaviour of the remote sensing-based blue crop water use estimation for different herbaceous and woody crops, against the official dataset for irrigation water accounting at two water management scales (of a relative mean absolute error of 15.4 % in the case of the largest water user association and of 17.1 % in the case of the river basin water authorities’ own estimations). The proposed approach, which we call Remote Sensing-based Agricultural Water Accounting and Footprint, aims to provide reliable and accurate spatially and temporally distributed thematic cartography about the remote sensing-based blue and green crop water use and water footprint. This information is essential for water managers with the goal of generating transparent and complementary information to incorporate into their own working scales.

Research on Leaf Area Index Inversion Based on LESS 3D Radiative Transfer Model and Machine Learning Algorithms

The Leaf Area Index (LAI) is a critical parameter that sheds light on the composition and function of forest ecosystems. Its efficient and rapid measurement is essential for simulating and estimating ecological activities such as vegetation productivity, water cycle, and carbon balance. In this study, we propose to combine high-resolution GF-6 2 m satellite images with the LESS three-dimensional RTM and employ different machine learning algorithms, including Random Forest, BP Neural Network, and XGBoost, to achieve LAI inversion for forest stands. By reconstructing real forest stand scenarios in the LESS model, we simulated reflectance data in blue, green, red, and near-infrared bands, as well as LAI data, and fused some real data as inputs to train the machine learning models. Subsequently, we used the remaining measured LAI data for validation and prediction to achieve LAI inversion. Among the three machine learning algorithms, Random Forest gave the highest performance, with an R2 of 0.6164 and an RMSE of 0.4109, while the BP Neural Network performed inefficiently (R2 = 0.4022, RMSE = 0.5407). Therefore, we ultimately employed the Random Forest algorithm to perform LAI inversion and generated LAI inversion spatial distribution maps, achieving an innovative, efficient, and reliable method for forest stand LAI inversion.

On the use of field RR Lyrae as Galactic probes

Context. The Vera C. Rubin Observatory will start operations in 2025. During its first two years, too few visits per target per band will be available, meaning that the mean magnitude measurements of variable stars will not be precise and thus standard candles such as RR Lyrae (RRL) will not be usable. Light curve templates (LCTs) can be adopted to estimate the mean magnitude of a variable star with a few magnitude measurements, provided that their period (plus the amplitude and reference epoch, depending on how the LCT is applied) is known. The LSST will provide precise RRL periods within the first six months, enabling exploitation of RRLs if LCTs are available. Aims. We aim to build LCTs in the LSST bands to enhance the early science with LSST. Using them will provide a one- to two-year advantage with respect to the classical approach concerning distance measurements. Methods. We collected grί-band data from the ZTF survey and z-band data from DECam to build the LCTs of RRLs. We also adopted synthetic grίz band data in the LSST system from pulsation models, plus SDSS, Gaia and OGLE photometry, inspecting the light amplitude ratios in different photometric systems to provide useful conversions to apply the LCTs. Results. We have built LCTs of RRLs in the grίz bands of the LSST photometric system; for the z band, we could build only fun damental mode RRL LCTs. We quantitatively demonstrated that LCTs built with ZTF and DECam data can be adopted on the LSST photometric system. The LCTs will decrease the uncertainty on distance estimates of RRLs by a factor of at least two with respect to a simple average of the available measurements. Finally, within our tests, we have found a brand new behavior of amplitude ratios in the Large Magellanic Cloud.

Electricity theft detection model based on the CEEMDAN-CNN-ViT

Abstract Most of the traditional power theft detection methods construct the model directly on the basis of the original power sequences, and do not simultaneously consider the long-period dependencies in the long-time sequences and the local connectivity features between periods, which limits the deep excavation of the behavioral laws of power users. In order to further improve the accuracy of power theft detection, this paper proposes a high-precision power theft detection model that integrates local anomaly filtering, energy consumption decomposition, and multi-feature fusion strategies. First, local anomaly filtering is used to eliminate local anomalies in the normalized energy consumption data to avoid the misleading effect of short-term abnormal behavior on the model. Then, the energy consumption decomposition based on CEEMDAN selects specific frequency band data that can accurately characterize the pattern of power theft users to improve the accuracy of power theft detection. Next, the long-time periodic features in the two-dimensional data and the short-time local features in the one-dimensional sequences are integrated by multi-feature fusion to enhance the adaptability of the model. The results show that the proposed model can effectively improve the detection accuracy, detection completeness, F1 score, and accuracy compared with the existing methods.

Exploring mass measurements of supermassive black holes in AGN using GAMA photometry and spectroscopy

Abstract In the era of massive photometric surveys, we explore several approaches to estimate the masses of supermassive black holes (SMBHs) in active galactic nuclei (AGN) from optical ground-based imaging, in each case comparing to the independent SMBH mass measurement obtained from spectroscopic data. We select a case-study sample of 28 type 1 AGN hosted by nearby galaxies from the Galaxy And Mass Assembly (GAMA) survey. We perform multi-component spectral decomposition, extract the AGN component and calculate the SMBH mass from the broad Hα emission line width and luminosity. The photometric g and i band data is decomposed into AGN+spheroid(+disc)(+bar) components with careful surface brightness fitting. From these, the SMBH mass is estimated using its relation with the spheroid Sérsic index or effective radius (both used for the first time on ground-based optical imaging of AGN); and the more widely used scaling relations based on bulge or galaxy stellar mass. We find no correlation between the Hα-derived SMBH masses and those based on the spheroid Sérsic index or effective radius, despite these being the most direct methods involving only one scaling relation. The bulge or galaxy stellar mass based methods both yield significant correlations, although with considerable scatter and, in the latter case, a systematic offset. We provide possible explanations for this and discuss the requirements, advantages and drawbacks of each method. These considerations will be useful to optimise stategies for upcoming high quality ground-based and space-borne sky surveys to estimate SMBH masses in large numbers of AGN.

Hyperspectral Data Can Differentiate Species and Cultivars of C3 and C4 Turf Despite Measurable Diurnal Variation

The ability to differentiate species is not adequate for modern forage breeding programs. The measurement of persistence is currently a bottleneck in the breeding system that limits the throughput of cultivars to the marketplace and prevents it from being selected as a trait. The use of hyperspectral data obtained through remote sensing offers the potential to reduce guesswork by identifying the distribution of pasture species, but only if such data alone can distinguish the subtle differences within species, i.e., cultivars. The implementation of this technology faces many challenges due to the spectral and temporal variability of species. To understand the spectral variability between and within species groups, differentiation using hyperspectral data from monoculture plots of turf species was utilized. Spectral data were collected over a year using an ASD FieldSpec® and canopy pasture probe (CAPP). The plots consisted of monocultures of various species, and cultivars (a total of 10 plots). Linear discriminant analysis (LDA) was conducted on the full spectrum and reduced band data. This technique successfully differentiated the species with high accuracy (>98%). We demonstrate the potential of hyperspectral data and analysis techniques to accurately separate differences down to cultivar level. We also show that diurnal variation is measurable in the spectra but does not preclude differentiation.

Correlation between Soil Moisture Change and Geological Disasters in E’bian Area (Sichuan, China)

E’bian Yi Autonomous County is a mineral-rich area located in a complex geological structure zone. The region experiences frequent geological disasters due to concentrated rainfall, steep terrain, and uneven vegetation cover. In particular, during the rainy season, large amounts of rainwater rapidly accumulate, increasing soil moisture and slope pressure, making landslides and debris flows more likely. Additionally, human activities such as mining, road construction, and building can alter the original geological structure, exacerbating the risk of geological disasters. According to publicly available data from the Leshan government, various types of geological disasters occurred in 2019, 2020, 2022, and 2023, resulting in economic losses and casualties. Although some studies have focused on geological disaster issues in E’bian, these studies are often limited to specific areas or types of disasters and lack comprehensive spatial and temporal analysis. Furthermore, due to constraints in technology, funding, and manpower, geophysical exploration, field geological exploration, and environmental ecological investigations have been challenging to carry out comprehensively, leading to insufficient and unsystematic data collection. To provide data support and monitoring for regional territorial spatial planning and geological disaster prevention and control, this paper proposes a new method to study the correlation between soil moisture changes and geological disasters. Six high-resolution Landsat remote sensing images were used as the main data sources to process the image band data, and terrain factors were extracted and classified using a digital elevation model (DEM). Meanwhile, a Normalized Difference Vegetation Index–Land Surface Temperature (NDVI-LST) feature space was constructed. The Temperature Vegetation Drought Index (TVDI) was calculated to analyze the variation trend and influencing factors of soil moisture in the study area. The research results showed that the variation in soil moisture in the study area was relatively stable, and the overall soil moisture content was high (0.18 < TVDI < 0.33). However, due to the large variation in topographic relief, it could provide power and be a source basis for geological disasters such as landslide and collapse, so the inversion value of TVDI was small. The minimum and maximum values of the correlation coefficient (R2) were 0.60 and 0.72, respectively, indicating that the surface water content was relatively large, which was in good agreement with the calculated results of vegetation coverage and conducive to the restoration of ecological stability. In general, based on the characteristics of remote sensing technology and the division of soil moisture critical values, the promoting and hindering effects of soil moisture on geological hazards can be accurately described, and the research results can provide effective guidance for the prevention and control of geological hazards in this region.

ATR-FTIR and NIR spectra combined with chemometrics and convolutional neural networks for identification of polygonati rhizome

Considering the high similarity among various rhizomes, the issue of adulteration and mixed use of polygonati rhizome (PR) in the market is widespread. Nowadays, the PR polysaccharides are the main pharmacological active ingredient. To ensure the uniformity of the product quality, it is essential to establish an effective and rapid method for PR identification. Herein, the polysaccharides content was determined via ATR-FTIR and NIR spectra of PR. The correlation between polysaccharides content and spectra were analyzed. Additionally, the partial least squares discriminant analysis (PLS-DA), support vector machines (SVM), and convolutional neural networks (CNN) were established using the data of ATR-FTIR spectra, NIR spectra, characteristic bands and feature variables. The obtained difference of the polysaccharides content was particularly large. By extracting latent variables and combining data fusion to establish PLS-DA, SVM and CNN, PR can be accurately identified. The content of PR polysaccharides can be predicted based on the absorbance according to ATR-FTIR and NIR spectra.

Solar bird banding: Notes on changes in avian behavior while mist-netting during an eclipse

Solar eclipses present rare celestial events that can elicit unique behavioral responses in animals, yet comprehensive studies on these phenomena, particularly concerning bird behavior, remain limited. This study, conducted at the Red Butte Canyon Research Natural Area in Utah during the annular solar eclipse on October 14, 2023, aimed to document and analyze avian activity using bird banding data. Leveraging 11 years of banding records, we observed a surprising positive peak in bird captures, indicating increased activity during the eclipse, challenging conventional expectations of decreased activity during peak totality. The unexpected, record-breaking captures on the eclipse day at this location, which also surpassed the average trend in captures over time for 18 other banding days in mid-October, highlights the complexity of bird behavior during celestial events. This study marks the first known published effort to conduct bird banding during a solar eclipse. Quantitative analyses, including species composition and capture trends, contribute to a nuanced understanding of avian responses to the eclipse. This study underscores the importance of empirical research in unraveling the intricacies of how birds navigate and adapt to unique environmental conditions created by solar eclipses.

Wildfire smoke impacts the body condition and capture rates of birds in California

ABSTRACT Despite the increased frequency with which wildfire smoke now blankets portions of world, the effects of smoke on wildlife, and birds in particular, are largely unknown. We used 2 decades of banding data from the San Francisco Bay Bird Observatory to investigate how fine particulate matter (PM2.5)—a major component and indicator of wildfire smoke—influenced capture rates and body condition of 21 passerine or near-passerine bird species. Across all study species, we found a negative effect of acute PM2.5 exposure and a positive effect of chronic PM2.5 exposure on avian capture rates. Together, these findings are indicative of decreased bird activity or local site removal during acute periods of wildfire smoke, but increased activity or site colonization under chronic smoke conditions. Importantly, we also observed a negative relationship between chronic PM2.5 exposure and body mass change in individuals with multiple captures per season. Our results indicate that wildfire smoke likely influences the health and behavior of birds, ultimately contributing to a shift in activity and body condition, with differential short-term versus long-term impacts. Although more research is needed on the mechanisms driving these observed changes in bird health and behavior, as well as validation of these relationships in more areas, our results suggest that wildfire smoke is a potentially frequent large-scale environmental stressor to birds that deserves increasing attention and recognition.

Random Forest Model-Based Inversion of Aerosol Vertical Profiles in China Using Orbiting Carbon Observatory-2 Oxygen A-Band Observations

Aerosol research is important for the protection of the ecological environment, the improvement of air quality, and as a response to climate change. In this study, a random forest (RF) estimation model of aerosol optical depth (AOD) and extinction coefficient vertical profiles was, respectively, established using Orbiting Carbon Observatory-2 (OCO-2) oxygen-A band (O2 A-band) data from China and its surrounding areas in 2016, combined with geographical information (longitude, latitude, and elevation) and viewing angle data. To address the high number of OCO-2 O2 A-band channels, principal component analysis (PCA) was employed for dimensionality reduction. The model was then applied to estimate the aerosol extinction coefficients for the region in 2017, and its validity was verified by comparing the estimated values with the Cloud-Aerosol Lidar Infrared Pathfinder Satellite Observations (CALIPSO) Level 2 extinction coefficients. In the comprehensive analysis of overall performance, an AOD model was initially constructed using variables, achieving a correlation coefficient (R) of 0.676. Subsequently, predictions for aerosol extinction coefficients were generated, revealing a satisfactory agreement between the predicted and the actual values in the vertical direction, with an R of 0.535 and a root mean square error (RMSE) of 0.107 km−1. Of the four seasons of the year, the model performs best in autumn (R = 0.557), while its performance was relatively lower in summer (R = 0.442). Height had a significant effect on the model, with both R and RMSE decreasing as height increased. Furthermore, the accuracy of aerosol profile inversion shows a dependence on AOD, with a better accuracy when AOD is less than 0.3 and RMSE can be less than 0.06 km−1.

A dynamic snow depth retrieval model based on time-series clustering optimization for GPS-IR

Due to the influence of environmental factors (i.e., terrain and surface coverage) around the GPS receivers, the snow depth retrieval results obtained by the existing global positioning system interferometric reflection (GPS-IR) method show significant variability. The resulting loss of reliability and accuracy limits the broad application of this technology. Therefore, this paper proposes a dynamic snow depth retrieval model based on time-series clustering optimization for GPS-IR to fully leverage multi-source satellite observation data for automatic and high-precision snow depth retrieval. The model employs Dynamic Time Warping distance measurement combined with the K-Medoids clustering algorithm to categorize frequency sequences obtained from various satellite trajectories, facilitating effective integration of multi-constellation data and acquisition of optimal datasets. Additionally, Long Short-Term Memory networks are integrated to capture and process the long-term dependencies in snow depth data, enhancing the model’s adaptability in handling time-series data. Validated against SNOTEL measured data and standard machine learning algorithms (such as BP Neural Networks, RBF, and SVM), the model’s retrieval capability is confirmed. For P351 and AB39 sites, the correlation coefficients for L1 band data retrieval were both 0.996, with RMSEs of 0.051 and 0.018 m, respectively. The experiment results show that the proposed model demonstrates superior precision and robustness in snow depth retrieval compared to the previous method. Then, we analyze the accuracy loss caused by sudden snowfall events. The proposed model and methodology offer new insights into the in-depth study of snow depth monitoring.