QLQingyang Liu

Qingyang Liu

*This is an algorithmically generated author record based on publicly available information.

Current Organizations / Institutes

The organizations/institutes from the scholar's latest article


Qingyang Liu has a total of 26 co-authors. Here are the top co-authors by number of shared publications:

  • Rui Li

    School of Digital Technology & Innovation Design, Jiangnan University, Wuxi, China

    9 shared publications
  • Peng Zhang

    Meteorological Observation Centre, China Meteorological Administration, Beijing 100081, China

    6 shared publications
  • Yipu Wang

    School of Earth and Space Sciences & State Key Laboratory of Fire Science & School of Information Science Technology University of Science and Technology of China Hefei China

    5 shared publications
  • Jiheng Hu

    4 shared publications
  • Jiheng Hu

    Department of Climate and Space Sciences and Engineering, University of Michigan, Ann Arbor, MI 48109, USA

    4 shared publications
  • Binbin Song

    School of Earth and Space Sciences & State Key Laboratory of Fire Science & School of Information Science Technology University of Science and Technology of China Hefei China

    3 shared publications

Latest Publications

  • Open Access

    Fire Danger and Intensity Response to Vegetation Properties and Weather Conditions in Eastern Asia Using Multilayer Perceptron

    by Qingyang Liu, Rui Li, Jiheng Hu, Dong Li, Hongwei Zhao, Jiawei Duan, Yuyun Fu

    While climate change has intensified wildfire danger, traditional prediction models primarily depend on weather-based indices and empirical thresholds, often neglecting the influence of time-varying fuel moisture—an essential factor in determining wildfire initiation and propagation. This limitation hinders a comprehensive understanding of fire-driving factors, particularly vegetation properties and their interactions with fire danger and intensity. Satellite-derived microwave emissivity difference vegetation index (EDVI) provides vegetation water content observations independent of soil moisture assumptions or empirical formulas. While prior studies have established EDVI fire relationships and employed EDVI for fire radiative power (FRP) estimation at ignition points, none have quantitatively integrated EDVI's role in pre-ignition fire likelihood and subsequent intensity modulation. To address this, we developed a data-driven two-stage MLP framework over eastern Asia (70 145°E, 10–55°N). The first stage is a binary classifier with residual connections and batch normalization for fire occurrence probability; the second stage applies square-root transformed FRP regression within burning areas. The fire danger model achieved an F1-score of 0.83 and AUC of 0.91, and the FRP regression model yielded an RMSE of 23.34 MW with a correlation of 0.45. Results demonstrate that integrating EDVI with optical indices and meteorological parameters significantly improves estimation of fire occurrence and intensity. Using Explainable AI (xAI) techniques, we revealed distinct regional drivers: EDVI exhibited heightened importance in arid North Asia, whereas Normalized Difference Vegetation Index (NDVI) dominated in temperate regions. This data-driven approach uncovers critical vegetation moisture-fire dynamics, laying the foundation for developing higher-precision fire spread models and advancing forecasting systems.

  • Open Access

    Downscaling Microwave‐Based Evapotranspiration With a Fourier‐Supervised Multi‐Source Fusion Network in Central‐Southern East Asia

    by Haoyang Li, Dong Li, Yipu Wang, Qingyang Liu, Jiheng Hu, Binbin Song, Shengli Wu, Peng Zhang, Danfeng Hong, Rui Li

    Evapotranspiration (ET) is a critical component of the land‐atmosphere energy and water cycle. Satellite remote sensing has proven to be highly effective for large‐scale ET estimation across heterogeneous landscapes, but producing high‐resolution, all‐weather ET remains difficult. Passive microwave remote sensing enables daily all‐weather ET retrievals, but suffers from coarse spatial resolution. Meanwhile, optical remote sensing provides finer spatial resolution but is hindered by prevailing cloud contamination, limiting its temporal availability. To leverage the complementary strengths of both data sources, we propose a Fourier‐supervised Multi‐source Fusion ET Downscaling Network (FMED‐Net), which fuses daily 0.25° microwave ET retrievals (ET_EDVI) and 8‐day 0.05° optical ET (ET_MODIS), along with various auxiliary data to produce all‐weather daily 0.05° ET_FMED. Integrating super‐resolution techniques and Fourier domain transformations, FMED‐Net effectively captures complex relationships in both the spatial and frequency domains across multi‐source data. The proposed method is applied to central‐southern East Asia during 2016–2018. Validations against in situ measurements at 10 eddy covariance sites suggest improved performance of 0.05° ET_FMED compared with the original 0.25° ET_EDVI (Nash–Sutcliffe Efficiency +30.27%, Kling–Gupta Efficiency +19.93%, Bias −37.43%). FMED‐Net also performs evidently better than the four representative machine learnings and deep learning models in both overall accuracy and the ability to preserve fine spatial structures. Furthermore, it effectively captures detailed spatial and temporal ET dynamics and maintains robust performance under cloudy conditions, compensating for the limitations of 8‐day optical ET products. These advantages highlight its potential for developing future global, daily, and all‐weather ET products based on multi‐source fusion.

  • Global Gross Primary Productivity Estimation Using Passive Microwave Observations From China's Fengyun‐3D Satellite

    by Binbin Song, Qingyang Liu, Jiheng Hu, Yipu Wang, Peng Zhang, Lin Chen, Shengli Wu, Rui Li

    In this study, we present the development and validation of a microwave‐based global gross primary productivity (GPP) estimation method, EDVI‐GPP, using the microwave emissivity difference vegetation index (EDVI) retrieved from the China's Fengyun‐3D satellite for the period 2020–2022. The global EDVI‐GPP model incorporates the effect of diffuse radiation from clouds on GPP and utilizes the Shuffled Complex Evolution‐University of Arizona (SCE‐UA) algorithm to tune its key parameters of light use efficiency and the seasonality of the vegetation water content index. In situ GPP measurements from 163 eddy covariance flux sites covering 10 major ecosystem types around the globe are used to calibrate and validate the model performance. EDVI‐GPP provides daily temporal resolution GPP estimations applicable under both clear and cloudy skies. At 8‐day temporal resolution, the coefficients of determination between EDVI‐GPP and in situ GPP measurements ( R 2 = 0.51) during 2020–2022 are comparable to several mainstream GPP products (FLUXCOM‐GPP, MODIS‐GPP, GLASS‐GPP, PML‐GPP, and GOSIF‐GPP), with a reduced bias of −0.46 gC m −2 day −1 . On a global scale, the annual‐averaged EDVI‐GPP exhibits high spatial consistency ( R = 0.71–0.74) with the compared GPP products. The EDVI‐GPP model quantifies the mean global GPP as 123.77 ± 1.33 Pg C yr −1 from 2020 to 2022, which is in close agreement with other published estimates. By using the three‐cornered hat (TCH) method to evaluate GPP uncertainty, we find that EDVI‐GPP exhibits a large uncertainty in agricultural areas. This research incorporates microwave‐derived variables into daily GPP estimation on a global scale, providing a less cloud‐affected and reliable measurement.

  • Global Evapotranspiration Retrieval Using Fengyun‐3D Passive Microwave Measurements With Genetic Algorithm Optimization

    by Qingyang Liu, Peng Zhang, Yipu Wang, Jiheng Hu, Rui Li

    Accurate estimation of terrestrial evapotranspiration (ET) is essential for understanding the global water cycle and energy balance. Satellite passive microwave (PMW) remote sensing offers a unique advantage for retrieving ET under cloudy conditions, where optical remote sensing fails. However, existing PMW‐based ET retrieval methods face uncertainties at a global scale, particularly across diverse land types and climates. Additionally, observations from China's Fengyun‐3D (FY‐3D) satellite have not been utilized for global ET estimation. To address these challenges, this study combined in situ ET measurements from 207 flux towers across 13 vegetation types with FY‐3D microwave radiation imager observations and reanalysis data sets. A machine‐learning genetic algorithm (GA) was developed to optimize key parameters of the PMW‐based ET retrieval method, generating a new global ET product (ETFY3D) with 0.25° resolution for 2020–2022. Evaluations at flux towers showed significantly improved accuracy compared to the original algorithm, reducing bias by 1.07 mm/day and root mean square error by 0.85 mm/day while increasing the Kling‐Gupta efficiency by 0.67. The ETFY3D was further evaluated against three independent global ET products: the Moderate Resolution Imaging Spectroradiometer product, the Global Land Evaporation Amsterdam Model, and the Penman‐Monteith‐Leuning (PML) product. The comparison showed that ETFY3D performed consistently with these benchmark products at a global scale. Sensitivity analysis highlighted vegetation and humidity parameters' importance. This study demonstrates AI's potential for optimizing physical‐based ET algorithms. And an independent new global ET product‐based China's FY low orbit satellite was developed for further applications.

  • Global Microwave Land Surface Emissivity Under all Weather Conditions Derived From Fengyun‐3D Satellite Observations

    by Qingyang Liu, Jiheng Hu, Peng Zhang, Yonghong Liu, Rui Li

    A global microwave land surface emissivity (MLSE) data set from January 2020 to December 2022 under all‐weather conditions was retrieved using microwave brightness temperatures from Fengyun‐3D Microwave Radiation Imager (MWRI), combined with cloud properties derived from the Medium Resolution Spectral Imager on the same satellite. The contributions of clouds and atmospheric particles were calculated using a four‐stream microwave radiative transfer model and subtracted from the Tbs measured at the top of the atmosphere. To evaluate the MWRI MLSE, we compared it with four reference clear‐sky emissivity data sets retrieved from Global Precipitation Mission Microwave Imager, Advanced Microwave Scanning Radiometer for EOS (AMSR‐E), and the Tool to Estimate Land Surface Emissivity at Microwaves and Millimeter waves (TELSEM) which is based on the measurements of Special Sensor Microwave/Imager (SSM/I). We also assessed the sensitivity of seasonal variations of MLSE to surface conditions indicated by the normalized difference vegetation index and precipitation seasonal cycles. Results demonstrate that our retrieval achieves good consistencies with the four reference products, exhibiting similar surface characteristics and spectral dependencies. Based on the all‐weather retrievals, we studied the potential impacts of clouds on MLSE in tropical regions. Results indicate that the presence of clouds may lead to a decrease in land surface emissivity due to potentially increased moisture in the soil‐vegetation mediums.


  • Monitoring Daily All‐Sky Evapotranspiration Over the East Asian Continent Using Multi‐Channel Passive Microwave Measurements From Fengyun‐3B Satellite of China

    by Yipu Wang, Jiheng Hu, Rui Li, Peng Zhang, Binbin Song, Qingyang Liu

    Terrestrial evapotranspiration (ET) modulates energy and water cycles of land‐atmosphere continuum. Satellite daily ET estimation under all sky (both clear and cloudy sky) remains challenging due to the difficulty in detecting vegetation dynamics under cloud cover. In this study, we utilized the microwave land surface Emissivity Difference Vegetation Index (EDVI) at X‐ and K‐bands, which was retrieved from Chinese Fengyun‐3B satellite Microwave Radiation Imager (MWRI) measurements under all sky conditions, to derive daily ET (ETFYEDVI) over East Asia (80–145°E, 10–55°N) from 2016 to 2018, in combination with multi‐source satellite all‐sky radiation and ERA5 climatic data. The Fengyun‐3B EDVI was used to represent the key vegetation characteristic for allocating total available energy and constraining canopy conductance. ETFYEDVI was validated at multiple spatiotemporal scales, against daily in situ measurements from eight flux sites and annual water balance method at nine river basins. Site‐scale validations showed that ETFYEDVI generated overall good accuracy (R2 = 0.78, bias of −0.1 mm day−1) across forest, grass, crop, and desert sites. Basin‐scale evaluations further demonstrated the stable and small bias (R2 of 0.96, bias of −22.2 mm year−1) in annual ETFYEDVI from southern humid to arid inland basins. This accuracy was comparable or better than those of four global satellite ET products. More importantly, ETFYEDVI produced low and stable accumulative bias under increasing cloud conditions, and characterized the spatiotemporal variability in continental ET even under heavy cloud cover (>90%) conditions. This study marks the first attempt to use Fengyun‐3 satellite passive microwave measurements to produce continental‐scale ET, showing great potential for mapping spatiotemporally continuous global daily ET.


  • Estimation of Fire Counts and Fire Radiative Power Using Satellite Optical and Microwave Vegetation Indices With Random Forest Method

    by Jiawei Duan, Jiheng Hu, Yuyun Fu, Qingyang Liu, Rui Li, Yipu Wang

    The satellite microwave emissivity difference vegetation index (EDVI) has been used in previous studies to estimate FCs and FRP using traditional multivariate linear regression models. However, the nonlinear effects and contributions of numerous factors that affect forest fires cannot be disentangled by this model. Using the random forest (RF) model, this study utilized multiple EDVIs and the optical normalized difference vegetation index (NDVI) as key fuel properties to resolve the physical driving mechanisms of forest fires and to estimate the daily FCs and FRP over East Asia. The results showed that the estimated FCs and FRP were in good agreement with satellite observations, with a spatial R of 0.59 for FCs and 0.63 for FRP and a temporal R of 0.80 for FCs and 0.81 for FRP. The integration of EDVIs and NDVI into the RF model was found to improve model performance and generate overall lower systematic errors than the model without vegetation variables. Model performance was better than that in previous studies using multivariate linear regression models. In addition, EDVIs showed greater importance than NDVI. This was largely due to their daily temporal resolution that allowed EDVIs to capture forest fire dynamics in time. The combination of the RF model with satellite microwave and optical observations shows good performance and has great potential for FC and FRP estimations in global fire danger assessment.


  • Open Access

    Simulations of Microwave Land Surface Emissivity Using FengYun-3D Microwave Radiation Imager Data: A Case in the Tibetan Plateau

    by Yonghong Liu, Fuzhong Weng, Fei Tang, Rui Li, Yongming Xu, Yang Han, Jun Yang, Qingyang Liu

    Accurate information on microwave land surface emissivity (MLSE) is important for satellite data assimilation. In this paper, a new random forest (RF) algorithm is developed for retrieving MLSE under all-sky conditions. Using Level-1 brightness temperature data from the FengYun-3D (FY-3D) Microwave Radiation Imager (MWRI) in 2022, two global MLSE daily product data sets, clear-sky (FY-3D1) and clear/cloudy (FY-3D2), were obtained by using one-dimensional variational method and microwave radiative transfer method, respectively. Based on the global spatiotemporal consistency assessment, a high-quality daily MLSE training dataset for the Tibetan Plateau was selected from the two data sets. Then, ten land surface parameters from routine observation were chosen as input features to the RF model, to simulate the MLSE under all-sky conditions in the Tibetan Plateau. The results show that both FY-3D1 and FY-3D2 MLSE data sets are comparable to the international mainstream MLSE products in quality, while the clear sky FY-3D1 is likely to be better than the clear/cloudy FY-3D2 MLSE. Land surface roughness, vegetation optical thickness, normalized vegetation index, and land cover type are identified as the most important factors affecting MLSE in the Tibetan Plateau. The RF model can effectively simulate the MLSE in the frequency range of 10.65∼89.0 GHz under all-sky conditions. The coefficients of determination (R2) for horizontal polarization and vertical polarization range from 0.86 (10.65GHz) to 0.91(18.7GHz) and from 0.60(10.65GHz) to 0.74(89.0GHz), respectively. The root-mean-square errors (RMSE) for horizontal polarization and vertical polarization range from 0.017 (23.8GHz) to 0.023(10.65GHz) and from 0.016 (10.65GHz) to 0.019(89.0GHz), respectively. These results indicate that machine learning is likely to be an effective method for future all-sky simulation of MLSE.