Significance The microphysical properties of cloud-precipitation systems play a critical role in atmospheric processes, weather prediction, and climate modeling. Characteristics such as the size, shape, and distribution of cloud and precipitation particles directly influence Earth's radiation budget, hydrological cycle, and precipitation dynamics. However, conventional radar observations primarily provide basic parameters, including reflectivity factors and optical thickness, which are insufficient for directly obtaining high-quality microphysical parameters. Mixed-phase clouds, where liquid water droplets and ice crystals coexist, exhibit complex microphysical structures and optical properties that significantly differ from those of purely liquid or purely ice clouds. Variations in ice fraction strongly affect the retrieval of cloud microphysical properties in the visible and near-infrared spectrum. Therefore, developing high-precision detection and retrieval techniques for cloud-precipitation microphysical parameters is of great scientific and practical significance, as it can improve the understanding of complex atmospheric processes and provide essential data for enhancing numerical weather prediction and climate models. Progress In recent years, the rapid advancement of optoelectronic technologies and microwave device fabrication has led to increasing applications of lidar, millimeter-wave cloud radar, and weather radar in probing vertical atmospheric structures, providing novel tools for studying cloud-precipitation microphysics. Various precipitation microphysical retrieval methods have matured, including raindrop size distribution retrieval, radar power spectrum analysis, and dual-or multi-frequency Doppler radar techniques. These methods allow relatively high-precision estimation of liquid water content, ice water content, mass-weighted particle diameters, and precipitation intensity. In contrast, radar-based detection of cloud microphysical parameters remains limited, particularly for mixed-phase clouds, where current multi-wavelength radar or lidar techniques are inadequate for accurate retrieval. Multisensor data fusion is recognized as a key approach to overcome these limitations. By integrating observations from multi-frequency radar, lidar, microwave radiometers, multi-spectral imagers, and ground-based disdrometers, it is possible to obtain a more comprehensive characterization of cloud and precipitation microphysical properties. For instance, the "Multidisciplinary drifting Observatory for the Study of Arctic Climate" (MOSAiC) expedition deployed various instruments onboard, providing high spatiotemporal resolution data to investigate the microphysics of mixed-phase clouds and precipitation. In terms of detection techniques, lidar offers high-precision measurements for small cloud droplets but has limited penetration capabilities. Millimeter-wave cloud radar is sensitive to medium-to-large particles but less effective for small cloud droplets. Terahertz (THz) radar, with wavelengths comparable to cloud particle sizes (0.1-3 mm), demonstrates higher sensitivity to fine cloud particles and light precipitation, making it suitable for probing microphysical details that are challenging for millimeter-wave radar. In particular, GGband radar (similar to 183 GHz) has shown advantages in retrieving microphysical properties of light rain and drizzle, outperforming traditional Wband radars. Nevertheless, THz radar faces significant challenges, including strong atmospheric attenuation due to water vapor and oxygen absorption, limited effective range, and technical difficulties in developing high-efficiency transmitters and high-sensitivity receivers. Conclusions and Prospects Overall, high-precision detection and retrieval of cloud-precipitation microphysical parameters form a fundamental basis for understanding atmospheric processes, improving weather forecasts, and refining climate simulations. Future research should focus on three main directions: 1) the advancement of multisensor collaborative observation techniques to integrate multi-frequency radar, lidar, microwave radiometers, and ground-based instruments, enhancing the accuracy of microphysical parameter retrieval for mixed-phase clouds and precipitation; 2) the development of THz cloud radar, addressing key challenges such as atmospheric attenuation compensation, efficient transmitter and receiver design, system integration, and miniaturization; 3) the integration of artificial intelligence and machine learning methods to improve retrieval algorithm precision and automation, particularly in capturing multi-scale and multi-phase cloud and precipitation particles. With these technological developments, it is expected that high-resolution and accurate microphysical data will be increasingly available, advancing the understanding of cloud-precipitation systems and promoting their application in meteorology, hydrology, and climate research.

An improved multi-feature fusion scheme is proposed to address the edge signal loss problem inherent in existing clutter filtering methods for Ka-band millimeter-wave cloud radar. A recognition model is first constructed based on the temporal and vertical continuity of the reflectivity factor of echo signals to perform preliminary clutter identification. Subsequently, morphological binary dilation operations are introduced to generate candidate regions along cloud and fog edges, and neighborhood analysis techniques are employed to achieve precise determination of signal boundaries. The proposed algorithm is validated against co-located lidar observations. Results demonstrate that the scheme effectively suppresses clutter while preserving cloud and fog edge signals with substantially improved completeness, thereby resolving the edge signal loss problem associated with existing clutter filtering approaches and enhancing the overall data quality of millimeter-wave cloud radar.

The melting layer (ML) is a key feature of stratiform precipitation systems and represents a critical transition region where solid hydrometeors undergo microphysical modifications associated with the transformation from ice and snow particles to liquid raindrops. The microphysical processes within the ML strongly influence precipitation phase, intensity, and spatial distribution, with important implications for improving numerical weather prediction (NWP), quantitative precipitation estimation (QPE), and weather modification operations. Non-polarimetric cloud radars generally identify the ML based on the vertical variations of reflectivity factor and radial velocity, referred to as the traditional Method-O. However, these methods exhibit limited capability in accurately determining the ML top height (MLTH) and are susceptible to misidentification under weak echo conditions or complex microphysical environments. To address these limitations, this study proposes a novel ML identification method based on Doppler spectral skewness (SK) using non-polarimetric millimeter-wave cloud radar observations. The proposed method exploits the distinct morphological evolution of Doppler power spectra associated with hydrometeor microphysical transitions and investigates the relationship between temperature and SK. The results demonstrate that the SK-based method provides enhanced sensitivity to early-stage spectral transitions within the upper ML transition region and enables more reliable identification of the upper and lower boundaries of the ML, including double-layer melting structures. Statistical analysis indicates that the proposed method identifies the upper transition boundary of the ML at an average temperature of approximately −2 °C, whereas the Method-O determines the corresponding temperature near 1.2 °C. The temperature difference suggests that significant modifications in hydrometeor size distribution and fall velocity may occur above the conventional ML, potentially associated with processes such as aggregation, particle growth, and early-stage phase modification. The maximum difference in the retrieved MLTH between the two methods reaches 0.67 km, while the ML bottom height (MLBH) retrievals show relatively small discrepancies.

During severe haze events, the boundary layer exhibits a complex vertical structure, while high aerosol loadings hinder high-resolution temperature and humidity measurements. To address this, a Raman-Mie lidar and retrieval algorithms for temperature, humidity, and aerosol optical properties were developed at Xi'an University of Technology, enabling high-resolution profiling of haze vertical structures. A 12 d haze episode was continuously monitored from formation to dissipation, providing detailed spatiotemporal variations of temperature, relative humidity, and aerosols. The boundaries of temperature inversion (TI) and aerosol layers were identified using a threshold method. The results revealed a strong coupling between aerosols and temperature during pollution evolution. Dome and stove effects were observed, with possible coexistence and interaction. The top of a decreasing-type aerosol layer formed a stratified dome-like structure that constrained vertical diffusion, with the temperature gradient of the elevated TI varying inversely with its depth. Both TI strength and humidity were strongly correlated with surface PM2.5 concentrations. Surface-based TI exhibited a clear diurnal variation, with TI peaks preceding aerosol peaks. The results suggest that strong elevated TI and weak turbulence in the lower layer may facilitate aerosol accumulation. Cloud layers not only suppress radiative heating but may also enhance near-surface humidity through virga processes, which may be conducive to increases in PM2.5 concentrations. During the dissipation stage, the rapid breakdown of TI and enhanced solar heating were critical for pollutant reduction, while efficient horizontal transport facilitated the complete clearance of aerosols within the boundary layer.

Clouds are a critical component of the Earth-atmosphere system, exerting strong influence on climate change and the global water cycle. Temperature is a fundamental parameter governing cloud development and evolution, and its vertical gradient directly shapes the vertical structure of clouds, thereby modulating key cloud physical processes. However, cloud temperature observations remain limited, and previous studies have primarily focused on single-layer clouds (SLC) or a single inversion layer. In this study, ten years of radiosonde data (2014-2023) in Xi'an were used to investigate the occurrence frequency and vertical temperature structure of SLC, double-layer clouds (DLC), and multi-layer clouds (MLC). The results show that the occurrence frequency of SLC tends to decrease over time, while the frequency of DLC and MLC increases year by year in Xi'an. Cloud top heights (CTH) exhibit significant seasonal variations. In SLC, temperature inversions (TI) occur more frequently in the upper cloud during winter and spring and are generally stronger at cloud tops than at cloud bases, with the probability of strong cloud-top TI exceeding that at the cloud base by more than a factor of two. While SLC is dominated by single-layer radiative cooling, TI in DLC and MLC exhibits complex interlayer radiative coupling mechanisms. In DLC and MLC, upper clouds tend to show shallow inversion depths under strong TI, whereas weak TI is associated with deeper inversions. Lower clouds, in contrast, exhibit nearly twice the probability of strong cloud-top TI compared with upper clouds. This pronounced enhancement suggests that radiative forcing from upper clouds plays a key intermediary role in strengthening TI in lower clouds. When temperature gradients are small, however, this enhancement is reduced and inversion development in lower clouds may be suppressed. Overall, these findings clarify the vertical coupling between cloud layers and TI, providing refined insights and valuable observational evidence for future studies on cloud-radiation interactions and cloud physical processes.

The vertical airflow velocity directly governs the growth and development of clouds. Accurately retrieving this velocity represents one of the pivotal and challenging aspects in cloud dynamics research. Unfortunately, no existing method can precisely retrieve the continuous vertical airflow velocity within clouds. To tackle this problem, this study assesses the feasibility of employing particles at the left end of the power spectrum as tracer particles by incorporating the intensity and critical threshold of tracer spectral points. Through an analysis of the impacts of turbulence, wind shear, and beam width on particles in various phases, the spectral width threshold is further established. Ultimately, the applicable conditions for the small particle tracing method are clearly defined. Additionally, based on power spectrum skewness and temperature, a method for retrieving vertical airflow velocity through spectral separation in a bimodal structure is proposed. The combination of these two methods can overcome the limitations inherent in single retrieval method and yield the vertical airflow velocity within the cloud's vertical structure under specific conditions.

This study develops a blue-band (458 nm) vertically-profiling Mie-scattering lidar based on the Scheimpflug principle incorporated into a conventional pulsed Raman-Mie lidar, extending the detection wavelength of prevailing atmospheric lidar techniques. A dedicated retrieval algorithm for the 458-nm aerosol backscattering coefficient has been established, and its associated errors have also been comprehensively quantified through simulation studies and practical atmospheric measurements. The algorithm utilizes boundary conditions and lidar ratios interpolated from 355-nm and 532-nm Raman retrievals to enable Fernald inversion of the 458-nm lidar signal. A simulation model, validated with airborne survey data, has been developed to thoroughly investigate the impact of key error sources. Atmospheric measurements under clear, dusty, foggy-hazy conditions have successfully demonstrated the feasibility of the combined lidar system and validated the performance of the proposed algorithm. It has been found out that the systematic error in the 458-nm aerosol backscattering coefficient induced by lidar ratio and boundary conditions are 8.5-11.5 % (mean), 3.8-24.9 % (mean) and 7.4-16.9 % (mean) for clear, dusty and foggy/hazy atmospheric conditions, respectively. The present work provides crucial theoretical foundations and technical support for enhancing the accuracy and reliability of lidarbased atmospheric aerosol measurements.

The hygroscopic growth of aerosols directly affects their particle size and optical properties, and plays a crucial role in the atmospheric transport and transformation processes, thus serving as an important basis for evaluating environmental and climatic impacts. However, experimental observations on the aerosol hygroscopicity growth characteristics have always been challenging. In this study, the parameter, aerosol fluorescence capacity, is introduced to explore its performance and characteristics in the aerosol hygroscopic growth. Based on a ground-based fluorescence-Raman-Mie lidar system developed by Xi'an University of Technology, the hygroscopic growth properties of aerosol fluorescence capacity are investigated by utilizing synchronous profiles of atmospheric temperature, humidity, backscattering coefficient, and fluorescence capacity. We are focused on the variation trend of the fluorescence capacity hygroscopic growth factor with relative humidity, and discussed its parameterized equation using the Hänel and Brock models, and the results are further compared with the traditional scattering hygroscopic growth factor. Two cloudy cases by lidar are involved in the study, and the results showed that, the fluorescence capacity hygroscopic growth factor can reach 1.82 and 1.60, while, the scattering hygroscopic growth factors were 1.13 and 1.46, respectively; The parameterized fitting results showed that, the Brock model achieved a fitting goodness of 0.94 and 0.98 for the fluorescence capacity hygroscopic growth factor, which were superior to the Hänel model's 0.83 and 0.86; The characteristic parameter κ for the fluorescence capacity hygroscopic growth factor by the Brock model were valued 0.25 and 0.21, significantly higher than those for the scattering hygroscopic growth factor, 0.11 and 0.14. In addition, a continuous lidar detection further indicated, the characteristic parameter κ for the fluorescence capacity hygroscopic growth factor of 0.41 was obtained, which was significantly higher than the κ of 0.19 for the scattered hygroscopic growth factor. The results demonstrated that, compared to the traditional scattering hygroscopic growth factor, the fluorescence capacity hygroscopic growth factor exhibits stronger sensitivity and characterization ability, and can be used as an important optical parameter for evaluating aerosol hygroscopic growth characteristics.

This study utilizes Ka-band millimeter-wave cloud radar (MMCR), assisted by a precipitation phenomenon instrument, to conduct case studies and analyses of convective precipitation, cumulus precipitation, and stratus precipitation in the Xi’an region. Using the Doppler spectral data of the MMCR, dynamic parameters such as vertical air motion velocity (updraft and downdraft) and particle terminal fall velocity within these three types of cloud precipitation were retrieved. The results show that above the melting layer, the maximum updraft velocity in convective clouds reaches 15 m·s−1, and the strong updraft drives cloud droplets to move upward at an average velocity of about 5 m·s−1. The average updraft velocity in cumulus clouds is greater than that in stratus clouds, with updrafts in cumulus and stratus mainly distributed within 1.5–3 m·s−1 and 1–2 m·s−1, respectively. The reflectivity factor of precipitation particles (Ze) is used to correct the equivalent reflectivity factor (Ka-Ze) after attenuation correction below the MMCR melting layer. The accuracy of calculating the raindrop concentration using the Ka-Ze of MMCR was improved below the melting layer. Based on the relationship between terminal fall velocity and particle diameter and using the conversion between the MMCR power spectrum and raindrop spectrum, the concentration, fall velocity, and particle diameter of raindrops are calculated below the melting layer. The results show that the average reflectivity factor, average concentration, and average particle diameter of raindrops follow the order of convective precipitation > cumulus precipitation > stratiform precipitation. However, the average terminal fall velocity distribution of raindrop particles follows a different order: convective precipitation > stratiform precipitation > cumulus precipitation.

An autocollimator is a popular angle measuring apparatus which lacks the capability to measure the roll angle. This paper proposes a novel roll angle sensor with a large measuring range that is based on the autocollimation principle. A modified right-angle prism (MRP) functions as a reflector to admit a collimated beam and return two outgoing beams to the sensor head. The roll angle of the MRP can be attained by analyzing the moving tracks of the two light spots focused on a photodetector. The mathematical model is derived in detail, and the experimental results show that the measuring accuracy of the proposed sensor is ±13.85 arcsec over a range of 360°. These results verify the feasibility of the proposed sensor for roll angle measurements that require a large measuring range.

In response to the limitations of traditional threshold increment methods in dynamic glare evaluation, this study integrates human visual characteristics with the adaptive properties of intelligent materials to investigate the impact of motion speed on dynamic vision. A fuzzy circle model is employed to simulate the human eye's refractive effect, analyzing the response of intelligent materials to variations in equivalent luminous screen brightness under different dynamic vision conditions. Dynamic vision detection and road lighting measurement experiments were conducted to validate the proposed approach. Based on these findings, a dynamic glare evaluation model incorporating the sensing and actuation mechanisms of smart materials was developed, enabling adaptive glare perception optimization in response to environmental changes. Experimental results indicate that the relative error between simulated and actual fuzzy circles is below 1%, while the deviation in the dynamic-to-static brightness ratio is only 0.4% in a stationary state, confirming the model's accuracy and reliability. Additionally, the model exhibits consistency with traditional static glare evaluation methods. This study provides a new theoretical foundation and practical framework for applying intelligent materials in dynamic visual perception assessment.

Considering that ozone is essential to understanding air quality and climate change, this study presents a deep learning method for predicting atmospheric ozone concentrations. The method combines an attention mechanism with a convolutional neural network (CNN) and long short-term memory (LSTM) network to address the nonlinear nature of multivariate time-series data. It employs CNN and LSTM to extract features from short time series, enhanced by the attention mechanism to improved short-term prediction accuracy. It takes eight meteorological and environmental parameters from 16,806 records (2018-2019) as input, which are selected principal component analysis (PCA). It features an attention-based CNN-LSTM hybrid deep learning model with specific settings: a time step of 5, a batch size of 25, 15 units in the LSTM layer, the ReLU activation function, 25 epochs, and an overfitting avoidance strategy with a dropout rate of 0.15. Experimental results demonstrate that this hybrid model outperforms individual models and the CNN-LSTM model, especially in forward prediction with a multi-hour time lag. The model exhibits a high coefficient of determination (R2 = 0.971) and a root mean square error of 3.59 for a 1-hour time lag. It also exhibits consistent accuracy across different seasons, highlighting its robustness and superior time-series prediction capabilities for ozone concentrations.

Raman lidar can achieve high spatial and temporal resolution retrieval of atmospheric water vapour vertical profiles. However, it is difficult to effectively solve the problem of limited daytime water vapour retrieval distances owing to the influence of the solar background light. To enhance the daytime water vapour retrieval capability of lidar, this paper proposes a technique for retrieving the water vapour vertical profile by integrating lidar and ground meteorological parameters based on the backpropagation neural network algorithm. This study constructed a neural network training model, maximized its retrieval accuracy, and achieve the retrieval of daytime water vapour profiles. Under strong background light conditions at noon in summer, the proposed method increases the maximum retrieval height by 2.5 km compared to traditional lidar retrieval methods. A regression analysis was conducted between the neural network retrieval method proposed in this study and traditional lidar retrieval methods within the effective daytime detection height. The results demonstrate that the proposed method exhibits high accuracy, achieving a correlation coefficient, a coefficient of determination, and a root mean square error of 0.951, 0.904, and 0.889 gkg- 1, respectively.

Objective Lidar,a high-performance active sensing technology,has gained significant traction in recent years across fields such as meteorology,climate science,and environmental monitoring.It serves as a principal methodology for detecting atmospheric physical properties by analyzing echo signals generated from interactions between emitted narrow-pulse lasers and atmospheric constituents,offering high resolution and extended detection range.The quadratic attenuation of lidar echo signals with distance causes single-pulse returns to be overwhelmed by ambient noise,necessitating multiple accumulations(time integration)to enhance the signal-to-noise ratio(SNR).However,conventional lidar data acquisition and integration methods involve onboard storage of data chains collected from a single trigger event.After reaching the predetermined number of accumulations,the stored data are sequentially read out and averaged.This approach introduces a time overhead during data readout,as data acquisition cannot occur simultaneously with readout,creating an acquisition dead time.When the repetition frequency of the lidar system increases to the kHz level,the interplay between acquisition and readout becomes a limitation,resulting in extended dead time and potential pulse omissions. Methods In this paper,a"read-accumulate-store"intellectual property(IP)was proposed,which was developed within a field-programmable gate array(FPGA)with the dual-port RAM and an adder,enabling temporal integration of corresponding points in a data linked list.Its core innovation lies in concurrent acquisition and integration:during each acquisition cycle,prior results are retrieved,accumulated with current data,and stored iteratively until the preset accumulation count is achieved.The temporal integration IP architecture,implemented in FPGA,comprises the components of photoelectric conversion module(transforms optical signals into electrical signals),analog-to-digital(A/D)module,accumulation(ADD)module,dual-port RAM storage and control unit.To achieve the designed function,the control unit executes the following steps.First,prior to the first trigger,all storage units in the linked list are reset to zero.Second,the first data point is acquired when the first trigger arrives,and the value stored in the first position of the linked list is read.These two values are summed and stored back in the first position.This process repeats N times to complete the sampling for the first trigger.Since the storage units were cleared before the first trigger,each unit contains the result of a single acquisition after this step.For the second trigger,the same procedure as in the previous step is repeated.As the values in each storage unit of the linked list are read during this acquisition,each unit now contains the cumulative result of the corresponding points from the first and second acquisitions after completion.For the M-th trigger,this process continues,with each storage unit holding the cumulative result of the corresponding points from the previous M acquisitions.Finally,once the predetermined number of accumulations,P,is reached,the accumulated data are read out and transmitted to the host computer,yielding the final integrated results.This method ensures efficient data processing by integrating acquisition,accumulation,and storage,thereby facilitating high-fidelity temporal integration for lidar systems. Results and Discussions To verify the effectiveness of the proposed design,a square wave signal,superimposed with a 0.2 V Gaussian white noise,with a period of 100 Hz and an amplitude of 1 V,was used as the input signal for testing.The results demonstrate that the signal becomes progressively smoother with the increase in accumulations,indicating significant noise signal attenuation(Fig.5).The SNR increased by 42 dB after 1000 accumulations,confirming that the data acquisition and integration module can achieve multiple acquisition accumulation to reduce background noise and improve SNR.For further verification,this module was compared with the DPO5104 digital oscilloscope of Tektronix and the PXI-9826 acquisition card of ADLINK Technology in actual measurements of laser radar echo signals.After correcting the signals obtained by the three devices by the square of the distance,The RSCS,which eliminates the factor of attenuation factor of light transmission,revealed that despite some variations,the waveform trends were fundamentally consistent,and the position data of the thin cloud layer measurements were largely concordant.Finally,this method was applied to a high-repetition-frequency polarized Mie lidar system,with a 5 kHz repetition frequency,achieving data acquisition at a 50 MHz sampling rate and performing over 80000 cumulative averages,successfully determining the extinction and depolarization ratio coefficients(Fig.8). Conclusions A distinctive requirement in the digitization of lidar echo signals,setting it apart from other methods,is the need for temporal integration.While current acquisition techniques effectively handle data collection and temporal integration at low repetition frequencies,limited research addresses these processes at lidar repetition frequencies in the kHz range.This paper presents a novel"read-accumulate-store"method that enables temporal integration of corresponding data points within a linked list structure.This approach simultaneously reads previous acquisition results,accumulates them with current acquisition data,and stores the resulting sum,achieving seamless temporal integration.To implement this method,an intellectual property architecture for temporal integration was developed within a FPGA,utilizing dual-port RAM and an adder.SNR analyses and practical testing demonstrate that this method enables high-speed data acquisition and temporal integration in high-repetition-frequency lidar systems.Additionally,the method offers flexible configuration of parameters,as channels,sampling frequency,sampling length,and integration settings can be adjusted by the hardware description language.Its single-chip integration capability enhances both cost-effectiveness and compactness.Given its versatility and performance,this method shows potential for standardization as a modular component in lidar systems,promoting widespread adoption in advanced sensing applications.

Objective Hyperspectral remote sensing technology has become increasingly crucial in agricultural and environmental monitoring. Targeting the high-dimensional spatial characteristics of hyperspectral remote sensing data, we propose an advanced machine learning method to extract and validate the characteristic bands associated with soil nutrient contents. By applying this method, characteristic bands of various soil nutrients can be identified, and their contents efficiently assessed. This research provides a solid theoretical foundation for large-scale, rapid soil nutrient monitoring using drone-based hyperspectral remote sensing. It presents a reliable solution for the development and application of hyperspectral remote sensing technology, offering significant value for agricultural production and environmental protection. Methods Spectral experiments are conducted using drones equipped with hyperspectral sensors, collecting soil reflectance data across 176 spectral bands within the range of 398-1003 nm. The extraction of soil nutrient characteristic bands involves two main steps. The importance of the 176 spectral bands is ranked using a combination of random forest (RF) and differential evolution (DE) algorithms. The random forest method evaluates the importance of each spectral band, while the differential evolution algorithm refines the selection of spectral features, ensuring that the most informative bands are retained. This process results in a subset of spectral features indicative of soil nutrient content. The analytic hierarchy process (AHP) is employed to determine the relative importance of the spectral features in the subset. By ranking and applying weight thresholds, characteristic bands of different soil nutrients are successfully identified from the hyperspectral data. Finally, a quantitative inversion model for soil nutrient content is developed using a back-propagation neural network (BPNN). Results and Discussions Using available potassium as an example, the extracted characteristic bands are identified as 469.0, 501.6, 581.2, 697.2, 791.8, 795.4, 802.5, and 954.8 nm. Among these, the three most significant bands, 469.0, 501.6, and 954.8 nm, show the highest correlation with soil potassium content, making them critical for accurate nutrient assessment. The back-propagation neural network model trained with these characteristic bands achieves remarkable results. In the Training set, the coefficient of determination (R-2) is 0.954 (Fig. 8), the ratio of performance to deviation (RPD) is 4.78, and the root mean square error (RMSE) is 14.32 mg/kg (Fig. 11). In the validation set, the model achieves R-2 of 0.848, RPD of 2.21, and RMSE of 16.71 mg/kg (Fig. 11). These results significantly outperform those obtained using the traditional first-order derivative mathematical transformation method, which yields R-2 values of 0.729 and 0.521, RPDs of 1.81 and 1.13, and RMSEs of 36.02 mg/kg and 191.05 mg/kg in the modeling and validation sets, respectively (Table 2). Conclusions The findings of this paper demonstrate the effectiveness and feasibility of machine learning methods for extracting hyperspectral soil nutrient characteristic bands. By integrating advanced algorithms such as random forest, differential evolution, and analytic hierarchy process, we offer a robust solution for the application of hyperspectral remote sensing technology. It enables more accurate and efficient soil nutrient assessments, significantly reducing the time and cost associated with traditional soil sampling and analysis. The successful extraction of characteristic bands and the development of a reliable predictive model underscore the potential of drone-based hyperspectral remote sensing technology for large-scale, rapid soil nutrient monitoring. This approach not only improves the precision of soil nutrient assessments, but also supports informed decision-making in agricultural production and environmental management. The outcomes contribute to enhanced crop yields, better resource allocation, and more sustainable agricultural practices.

Objective Meteorological factors have a significant effect on the vertical distribution of aerosols,and the formation,accumulation,and dissipation of heavy pollution processes in winter are usually controlled by meteorological conditions.There is a high correlation between meteorological factors and the vertical structure of aerosols.Obtaining detailed evolution characteristics of meteorological factors is of great research value for studying the process of haze generation and dissipation.Among various meteorological factors,the vertical distribution of temperature plays a crucial role in the aggregation of aerosols in the boundary layer.Currently,ground meteorological stations,meteorological satellites,radiosondes,and microwave radiometers are the main means of detecting vertical temperature profiles,but none of them can achieve high spatiotemporal resolution detection of temperature profiles within the boundary layer.Rotational Raman lidar,as an effective technique for atmospheric temperature measurement,offers high temporal and spatial resolution,which is advantageous for studying atmospheric physical processes within the boundary layer during haze conditions.However,the system's detection performance within the boundary layer is significantly affected by the inconsistency of the bottom detection blind zone and signal attenuation caused by aerosols within the boundary layer,as well as interference from elastic scattering.To enable atmospheric temperature measurements in the bottom layer of haze conditions,we propose a temperature correction technique based on the backscatter ratio.We hope to effectively obtain the vertical structure of atmospheric temperature within the boundary layer of haze weather through this technique,thereby providing data support for the refined study of atmospheric physical processes under haze conditions. Methods The core of this technique involves constructing a linear functional relationship between the backscatter ratio and the elastic scattering crosstalk ratio and using the backscatter ratio to correct the rotational Raman ratio.The consistency of the geometric overlap factor of the rotational Raman channels significantly affects the bottom detection performance of the lidar.Therefore,accurately obtaining the ratio of high and low quantum number rotational Raman channels is a prerequisite for implementing rotational Raman temperature measurements.First,we use experimental data under clear sky conditions without haze to calibrate the geometric overlap factor ratio of the rotational Raman channel signals.Then,we calibrate the inversion function using a radiosonde under simultaneous spatial conditions and obtain the theoretical Raman ratio within the haze layer based on the temperature data from the radiosonde.Based on this theoretical ratio and the rotational Raman ratio that includes elastic scattering crosstalk,we calculate the elastic scattering crosstalk ratio.We perform a linear regression analysis on both the backscatter ratio and the elastic scattering crosstalk ratio to derive the corresponding system calibration constant.Finally,using this calibration constant and the measured backscatter ratio,we complete the correction of the rotational Raman ratio,allowing for the retrieval of the true atmospheric temperature data in the elastic scattering region. Results and Discussions Numerical simulation results indicate that inconsistencies in the geometric overlap factor and the elastic scattering crosstalk can lead to retrieval biases in atmospheric temperature measurements within the boundary layer during haze events.After applying dual corrections for the geometric overlap factor and the rotational Raman ratio,the temperature retrieval bias is reduced to less than 0.2 K(Fig.4).Experimental results show that the corrected temperature profile from the morning of of December 26,2023 exhibits a high degree of consistency with the radiosonde data obtained under simultaneous spatial conditions,with a maximum temperature bias of less than 1.4 K,while the uncorrected temperature bias is up to 7 K(Fig.8).On the evening of of December 24,2023,the maximum temperature bias is less than 0.6 K,while the uncorrected temperature bias is approximately 4 K(Fig.9).Additionally,continuous observation results clearly illustrate the vertical distribution of the temperature field.A comparison of the backscatter ratios of aerosols reveals a close correlation between the inversion temperature features and the vertical distribution of aerosols(Fig.10). Conclusions In our study,we propose a temperature correction technique based on the backscatter ratio to achieve atmospheric temperature measurements within the bottom layer of haze conditions.Simulation and experimental results indicate that this technique can effectively achieve precise measurements of atmospheric temperature within the boundary layer during haze events,clearly illustrating the vertical structure of atmospheric temperature and the characteristics of inversion.Since the fundamental basis for temperature correction relies on the correlation between the elastic scattering crosstalk ratio and the backscatter ratio,it is crucial to accurately obtain the backscatter ratio and the rotational Raman ratio within the haze layer.Therefore,highly consistent preprocessing of the lidar echo signals is necessary.Additionally,the stability of parameters such as laser energy,the geometric overlap relationship of the light transmission and reception system,the elastic scattering suppression ratio of the rotational Raman channel,and the photoelectric conversion efficiency also influence the implementation of this technique.Variations in these parameters can directly affect system stability,thereby affecting the temperature correction results.Therefore,a lidar system with high stability is an essential prerequisite for conducting atmospheric temperature corrections.Any minor adjustments to system parameters necessitate a reevaluation of the system calibration.In summary,the introduction of this correction technique provides scientific data and technical means for studying atmospheric physical processes within the boundary layer during haze events,facilitating a detailed investigation and analysis of the formation and evolution characteristics of haze.

Natural seeder-feeder is a common phenomenon in stratiform clouds precipitation. The quantitative evaluation of the enhancement effect of seeder-feeder plays an important role in improving the accuracy of weather forecasting and artificial weather operations. In this study, ground-based millimeter wave cloud radar, microwave radiometer, and ground precipitation phenomenon instrument were used to observe the natural seeder-feeder process between double-layer clouds during ground snowfall in the Xi'an, China for 3 years. A constrained small particle tracing method and an optimized fuzzy logic algorithm were proposed to achieve inversion of vertical air motions ( Va ) and particle terminal falling velocity ( Vf ), and identify cloud phase states. The reflectivity factor ( Z ) (attenuation corrected) and mean Doppler velocity ( Vr ) were used to identify the feeding and non-feeding areas in low-level feeding clouds. The enhancement effect of seeding on cloud particles was quantitatively evaluated by analyzing the differences in Z, Vf , and particle diameter ( D ) between the feeding and non-feeding areas. Meanwhile, the enhancement effect of seeding on ground precipitation was analyzed. The statistical results show that after seeding, both cloud particles and ground precipitation significantly increased. The average- Z , - Vf , and - D of cloud particles in the feeding areas have increased by 10 dBZ, 0.75 m·s -1 , and 0.5 mm compared to the non-feeding areas, respectively. Compared with the non-feeding period, the average ice water content (IWC), average number concentration ( NT ), mean particle diameter ( D m ), and maximum particle diameter (MPD) of ground precipitation particles during the feeding periods were [1.01-67.91], [1.04-35.37], [1.04-2.04], and [1.01-3.65] times, respectively.

An autocollimator is a goniometer established according to the principle of autocollimation, but it is ineffective for measuring the roll angle. This paper proposes an improved autocollimator available for large-range roll angle measurement, which maintains the optical structure of the classic one while incorporating a wedge prism (WP) working in transmissive mode as a roll angle sensing element to dissociate the collimated beam into two beams. According to the moving paths of the two light spots focused on the photodetector, the roll angle of the WP can be solved. The measuring method is expounded, and the calibration results reveal that the improved autocollimator has an accuracy of ±13.55 arcsec over a range of 360°, confirming its feasibility for roll angle measurement where a large measuring range is required.

In order to monitor forest fires, a high-repetition-rate polarization Lidar system was developed based on the light scattering and polarization characteristics of smoke particles generated during fires. The system consists of subsystems for laser emission, optical reception, echo signal acquisition and processing, and scanning control. To meet the demands for large-scale, high-resolution, and rapid forest fire detection, a high-power, high-repetition-rate laser was selected as the probing source, coupled with a high-resolution gimbal for precise scanning. With a Lidar repetition rate of 5 kHz, the system can perform patrol scanning a forest area with a 10-km radius in 48 minutes at an angular resolution of 1 degrees. To address the challenges of echo signal acquisition and cumulative averaging during high-repetition-rate detection, a novel "readout-accumulation-storage" IP (Intellectual Property) architecture was designed, enabling efficient echo signal processing and improving the signal-to-noise ratio. The completed high-repetition-rate polarization Lidar underwent near-field and far-field simulation experiments, with detected signal peaks corresponding to fire locations. When deployed in Yan'an City, the Lidar successfully detected simulated fires at distances of 5.4 km and 8. 1 km, validating the system's effective detection capability.

The roll angle is an important albeit difficult-to-measure geometric parameter. This study proposes a roll angle measurement method for an entire circle (360°). An elaborately designed double-sided optical wedge (DOW) is employed as a roll angle probe to divide the incident light beam into two beams. Both beams are imaged using a photodetector, and the roll angle of the DOW is obtained through the detection of the motion trajectories of the two images. The measurement principle is comprehensively analyzed, and the experimental results indicate that the accuracy of the constructed measurement system is better than ±7.72arcsec in the range of 360°. The proposed method is confirmed to be effective for large-range roll angle measurements.