Accurate and reliable monitoring of biomass in tropical forest has been a challenging task because a large proportion of forest is inaccessible. For effective implementation of REDD-plus and fair benefit sharing, monitoring methodology should be based on scientifically robust estimation of sources and sinks to meet MRV requirements. Though there have been major advances in satellite remote sensing technologies in recent years, none of them have been able to overcome the saturation problem that makes it hard to detect forests with high above-ground biomass volume and assess degradation. The saturation problem in biomass estimation can be overcome by adopting airborne LiDAR, because laser pulses penetrate even through a dense multi-layered canopy and there is a strong correlation between LiDAR data and biomass. Integrating different remote sensing and field reference data provides an accurate, precise, and affordable monitoring solution for tropical forests. In this regard, a two-phase sampling scheme optimizes field data collection efforts for model calibration and assists with objective and efficient positioning of sample plots. In the second sampling phase, estimates for a set of LiDAR transects are available, and the radiometric properties of satellite imagery are applied to identify the best estimators for target variables. Such a method is proposed here. It integrates sample plots with LiDAR transects and satellite images, and it attains a relative RMSE of 25 to 35 percent in aboveground biomass already on an area of 0.5 ha. Alternative methods, such as National Forest Inventories based on permanent sample plots, optical satellite imagery, and k-NN estimation or visual inspection, attain this error level only on areas of 100 ha or even more. Such high spatial resolution is crucial for awarding REDD credits to local forest owners. Arbonaut Ltd. has developed a forest inventory process and user-friendly tools (ArboLiDAR) to estimate above- and below-ground carbon stocks. The estimation process relies on a unified Bayesian statistical methodology, making it possible to incorporate various information sources such as direct measurements, quantities interpreted from remote sensing, and the results of modelling of carbon sinks such as below-ground carbon. These estimates can also be simply updated whenever new data becomes available.