Research progress on uncertainty in remote sensing-based forest reserves estimation

The three forest reserves, namely forest volume, biomass and carbon storage, are important indicators for monitoring forest resources at multiple levels, and also important parameters reflecting the quality and productivity of forest ecosystems. Remote sensing estimation of forest reserves is not a direct measurement process, but rather relies on the predictive model established between remote sensing variables and field-measured forest reserves. It inherently involves many sources of error and uncertainty, including sampling constraints, measurement inaccuracies, sensor noise, resolution constraints, atmospheric condition variations, model selection and parameter estimation uncertainty. Systematically identifying, quantifying and controlling these errors is crucial for enhancing the accuracy and reliability of remote sensing estimation of forest reserves. Therefore, this paper analyzes uncertainties in remote sensing-based estimation of forest reserves from three aspects: forest inventory samples, remote sensing data sources, and estimation models. Furthermore, it discusses and prospects methods for error control and uncertainty quantification in forest reserves estimation. This study is helpful to deeply understand the uncertainty sources in remote sensing estimation of forest reserves, and can provide a reference for improving the precision of forest reserves retrieval, as well as for optimizing strategies for forest resource monitoring and management.