Estimation of forest aboveground biomass (AGB) is crucial for various technical and scientific applications, ranging from regional carbon and bioenergy policies to sustainable forest management. However, passive optical remote sensing, which is the most widely used remote sensing data for retrieving vegetation parameters, is constrained by spectral saturation problems and cloud cover. On the other hand, LiDAR data, which have been extensively used to estimate forest structure attributes, cannot provide sufficient spectral information of vegetation canopies. Thus, this study aimed to develop a novel synergistic approach to estimating biomass by integrating LiDAR data with Landsat 8 imagery through a deep learning-based workflow. First the relationships between biomass and spectral vegetation indices (SVIs) and LiDAR metrics were separately investigated. Next, two groups of combined optical and LiDAR indices (i.e., COLI1 and COLI2) were designed and explored to identify their performances in biomass estimation. Finally, five prediction models, including K-nearest Neighbor, Random Forest, Support Vector Regression, the deep learning model, i.e., Stacked Sparse Autoencoder network (SSAE), and multiple stepwise linear regressions, were individually used to estimate biomass with input variables of different scenarios, i.e., (i) all the COLI1 (ACOLI1), (ii) all the COLI2 (ACOLI2), (iii) ACOLI1 and all the optical (AO) and LiDAR variables (AL), and (iv) ACOLI2, AO and AL. Results showed that univariate models with the combined optical and LiDAR indices as explanatory variables presented better modeling performance than those with either optical or LiDAR data alone, regardless of the combination mode. The SSAE model obtained the best performance compared to the other tested prediction algorithms for the forest biomass estimation. The best predictive accuracy was achieved by the SSAE model with inputs of combined optical and LiDAR variables (i.e., ACOLI1, AO and AL) that yielded an R² of 0.935, root mean squared error (RMSE) of 15.67 Mg/ha, and relative root mean squared error (RMSE_r) of 11.407%. It was concluded that the presented combined indices were simple and effective by integrating LiDAR-derived structure information with Landsat 8 spectral data for estimating forest biomass. Overall, the SSAE model with inputs of Landsat 8 and LiDAR integrated information resulted in accurate estimation of forest biomass. The presented modeling workflow will greatly facilitate future forest biomass estimation and carbon stock assessments.

Timely, spatially complete, and reliable forest above-ground biomass (AGB) data are a prerequisite to support forest management and policy formulation. Traditionally, forest AGB is spatially estimated by integrating satellite images, in particular, optical data, with field plots from forest inventory programs. However, field data are limited in remote and unmanaged areas. In addition, optical reflectance usually saturates at high-density biomass level and is subject to cloud contaminations. Thus, this study aimed to develop a deep learning based workflow for mapping forest AGB by integrating Landsat 8 and Sentinel-1A images with airborne light detection and ranging (LiDAR) data. A reference AGB map was derived from the wall-to-wall LiDAR data and field measurements. The LiDAR plots—stratified random samples of forest biomass extracted from the LiDAR simulated strips in the reference map—were adopted as a surrogate for traditional field plots. In addition to the deep learning model, i.e., stacked sparse autoencoder network (SSAE), five different prediction techniques including multiple stepwise linear regressions, K-nearest neighbor, support vector machine, back propagation neural networks, and random forest were individually used to establish the relationship between LiDAR-derived forest biomass and the satellite predictors. Optical variables (Landsat 8 OLI), SAR variables (Sentinel-1A), and their combined variables were individually input to the six prediction models. Results showed that the SSAE model had the best performance for the forest biomass estimation. The combined optical and microwave dataset as explanatory variables improved the modeling performance compared to either the optical-only or microwave-only data, regardless of prediction algorithms. The best mapping accuracy was obtained by the SSAE model with inputs of optical and microwave integrated metrics that yielded $R^{2}$ of 0.812, root mean squared error (RMSE) of 21.753 Mg/ha, and relative RMSE (RMSE$_{r}$) of 14.457%. Overall, the SSAE model with inputs of combined Landsat 8 OLI and Sentinel-1A information could result in accurate estimation of forest biomass by using the stratification-sampled and LiDAR-derived AGB as ground reference data. The modeling workflow has the potential to promote future forest growth monitoring and carbon stock assessment across large areas.

Estimation of forest aboveground biomass is critical for regional carbon policies and sustainable forest management. Passive optical remote sensing and active microwave remote sensing both play an important role in the monitoring of forest biomass. However, optical spectral reflectance is saturated in relatively dense vegetation areas, and microwave backscattering is significantly influenced by the underlying soil when the vegetation coverage is low. Both of these conditions decrease the estimation accuracy of forest biomass. A new optical and microwave integrated vegetation index (VI) was proposed based on observations from both field experiments and satellite (Landsat 8 Operational Land Imager (OLI) and RADARSAT-2) data. According to the difference in interaction between the multispectral reflectance and microwave backscattering signatures with biomass, the combined VI (COVI) was designed using the weighted optical optimized soil-adjusted vegetation index (OSAVI) and microwave horizontally transmitted and vertically received signal (HV) to overcome the disadvantages of both data types. The performance of the COVI was evaluated by comparison with those of the sole optical data, Synthetic Aperture Radar (SAR) data, and the simple combination of independent optical and SAR variables. The most accurate performance was obtained by the models based on the COVI and optical and microwave optimal variables excluding OSAVI and HV, in combination with a random forest algorithm and the largest number of reference samples. The results also revealed that the predictive accuracy depended highly on the statistical method and the number of sample units. The validation indicated that this integrated method of determining the new VI is a good synergistic way to combine both optical and microwave information for the accurate estimation of forest biomass.

Aboveground biomass (AGB) is a vital indicator for studying carbon sinks in forest ecosystems. Semiarid forests harbor substantial carbon storage but received little attention due to the high spatial–temporal heterogeneity that complicates the modeling of AGB in this environment. This study assessed the performance of different data sources (annual monthly time-series radar was Sentinel-1 [S1]; annual monthly time series optical was Sentinel-2 [S2]; and single-temporal airborne light detection and ranging [LiDAR]) and seven prediction approaches to map AGB in the semiarid forests on the border between Gansu and Qinghai Provinces in China. Five experiments were conducted using different data configurations from synthetic aperture radar backscatter, multispectral reflectance, LiDAR point cloud, and their derivatives (polarimetric combination indices, texture information, vegetation indices, biophysical features, and tree height- and canopy-related indices). The results showed that S2 acquired better prediction (coefficient of determination [R²]: 0.62–0.75; root mean square error [RMSE]: 30.08–38.83 Mg/ha) than S1 (R²: 0.24–0.45; RMSE: 47.36–56.51 Mg/ha). However, their integration further improved the results (R²: 0.65–0.78; RMSE: 28.68–35.92 Mg/ha). The addition of single-temporal LiDAR highlighted its structural importance in semiarid forests. The best mapping accuracy was achieved by XGBoost, with the metrics from the S2 and S1 time series and the LiDAR-based canopy height information being combined (R²: 0.87; RMSE: 21.63 Mg/ha; relative RMSE: 14.45%). Images obtained during the dry season were effective for AGB prediction. Tree-based models generally outperformed other models in semiarid forests. Sequential variable importance analysis indicated that the most important S1 metric to estimate AGB was the polarimetric combination indices sum, and the S2 metrics were associated with red-edge spectral regions. Meanwhile, the most important LiDAR metrics were related to height percentiles. Our methodology advocates for an economical, extensive, and precise AGB retrieval tailored for semiarid forests.

Accurate high-resolution soil moisture mapping is critical for surface studies as well as climate change research. Currently, regional soil moisture retrieval primarily focuses on a spatial resolution of 1 km, which is not able to provide effective information for environmental science research and agricultural water resource management. In this study, we developed a quantitative retrieval framework for high-resolution (250 m) regional soil moisture inversion based on machine learning, multisource data fusion, and in situ measurement data. Specifically, we used various data sources, including the normalized vegetation index, surface temperature, surface albedo, soil properties data, precipitation data, topographic data, and soil moisture products from passive microwave data assimilation as input parameters. The soil moisture products simulated based on ground model simulation were used as supplementary data of the in situ measurements, together with the measured data from the Maqu Observation Network as the training target value. The study was conducted in the Zoige region of the Tibetan Plateau during the nonfreezing period (May–October) from 2009 to 2018, using random forests for training. The random forest model had good accuracy, with a correlation coefficient of 0.885, a root mean square error of 0.024 m³/m³, and a bias of −0.004. The ground-measured soil moisture exhibited significant fluctuations, while the random forest prediction was more accurate and closely aligned with the field soil moisture compared to the soil moisture products based on ground model simulation. Our method generated results that were smoother, more stable, and with less noise, providing a more detailed spatial pattern of soil moisture. Based on the permutation importance method, we found that topographic factors such as slope and aspect, and soil properties such as silt and sand have significant impacts on soil moisture in the southeastern Tibetan Plateau. This highlights the importance of fine-scale topographic and soil property information for generating high-precision soil moisture data. From the perspective of inter-annual variation, the soil moisture in this area is generally high, showing a slow upward trend, with small spatial differences, and the annual average value fluctuates between 0.3741 m³/m³ and 0.3943 m³/m³. The intra-annual evolution indicates that the monthly mean average soil moisture has a large geographical variation and a small multi-year linear change rate. These findings can provide valuable insights and references for regional soil moisture research.

Hyperspectral images with very high resolution (VHR-HSI) have become considerably valuable due to their abundant spectral and spatial details. Classification of hyperspectral images (HSIs) is a basic and important procedure for diverse applications. However, low interclass spectral variability and high intraclass spectral variability in VHR-HSI, shadows, pedestrians, and low signal-to-noise ratio increase the fuzziness of different categories. To address the known challenges of VHR-HSI classification, an effective classification method based on encoder–decoder architecture is proposed. The proposed algorithm is an object-level contextual convolution neural network based on an improved residual network backbone with 3-D convolution, which fully considers the spatial–spectral and contextual features of HSIs. Two different spatial resolution aerial HSIs are used as experimental data. The results show that the overall accuracy of the proposed method is improved by 7.42% and 18.82%, respectively, compared to the pixelwise convolution neural network and DeepLabv3 algorithm, which is extraordinarily suitable for HSI classification with very high spatial resolution.