JOURNAL OF NANJING FORESTRY UNIVERSITY ›› 2023, Vol. 47 ›› Issue (1): 1-12.doi: 10.12302/j.issn.1000-2006.202209009

Previous Articles     Next Articles

A review of remote sensing application in national forest inventory

DING Xiangyuan(), CHEN Erxue(), LI Zengyuan, ZHAO Lei, LIU Qingwang, XU Kunpeng   

  1. Key Laboratory of Forestry Remote Sensing and Information System, National Forestry and Grassland Administration, Research Institute of Forest Resource Information Techniques, CAF, Beijing 100091, China
  • Received:2022-09-05 Accepted:2022-11-06 Online:2023-01-30 Published:2023-02-01
  • Contact: CHEN Erxue E-mail:dxy4201@126.com;chenerx@ifrit.ac.cn

Abstract:

National (continuous) forest inventory (NFI/NCFI) is an important part of forest resources monitoring system, which can provide timely and effective scientific basis for formulating national forestry development strategies and adjusting forestry policies(follow-up called NFI). Remote sensing has played a vital role in promoting the progress of NFI technology. It has become an indispensable technical means to support the operation of NFI. In terms of using remote sensing data as auxiliary information in NFI to increase the accuracy and efficiency of population parameter estimation, scholars at home and abroad have carried out a large number of studies on estimation models and methods, which can be summarized into four categories: design-based inference method, design-based and model-assisted method, model-dependent method, design and model hybrid method. Focusing on these four categories of estimation models and methods, this research summarizes the current research status at home and abroad, analyzes the problems existing in domestic related research, and gives suggestions on the follow-up key research and development directions and contents, hoping to promote the comprehensive application of Space-Air-Earth multi-source observation data in China’s NFI business. In terms of design-based inference method, there is little difference between domestic and foreign; a large number of design-based and model-assisted method research has been carried out abroad and has been applied to NFI business, but there are few domestic related research, business application is only embodied in area ratio estimation. Few research has been carried out on the estimation of quantitative forest parameters using the design-based and model-assisted method. The application, demonstration and promotion of this method should be strengthened in the future. The model-dependent method is the most basic method used by remote sensing in forest resources survey and monitoring. Lots of research has been done on the application of model method in NFI abroad, and the uncertainty measurement method in the collaborative application of multi-source data has been studied in depth. Also there are many domestic studies on model-dependent methods, but the systematic research on how to scientifically evaluate the fitting effect of the model and how to measure the uncertainty of the model estimation results are still need, which should be the focus of follow-up research. For the monitoring of forest resources in the area that difficult to investigate, the advantage of the model-dependent method, which is most conducive to solve the problem of small area estimation, should be fully utilized, and remote sensing as auxiliary data should be used to achieve effective estimation of forest parameters at different scales through the model-dependent method. Three types of design and model hybrid methods have been developed abroad for the application of NFI. There is little research on the first type of design and model hybrid method in China. The research on the second type of design and model hybrid method is limited to estimate land area by the double regression sampling method. Few research has been carried out on the design and model hybrid method of quantitative parameters such as stock volume; however, the research on the third type of design and model hybrid method has not yet adopted the idea of “data assimilation” to carry out relevant application research. It is recommended to strengthen the in-depth research and application demonstration of these three types of mixed methods in domestic NFI in the future.

Key words: national forest inventory, remote sensing applications, statistical inference, design-based inference method, design-based and model-assisted method, model-dependent method, design and model hybrid method

CLC Number: