不知道这个是不是你想要的。

Foreword to the Special Issue on “Recent Remote Sensing Applications”

The field of remote sensing is advancing rapidly. To summarize what has been achieved and provide outlooks for the future, the 9th International Symposium on Physical Measurements and Signature in Remote Sensing (ISPMSRS) was held at the Institute for Geographical Sciences and Natural Resource Research, Oct. 17-19, 2005, in Beijing, China. This successful meeting attracted more than 400 participants. Following the symposium, selected original research papers are being published in three special issues in Remote Sensing of Environment, Journal of Remote Sensing and Photogrammetric Engineering and Remote Sensing. The review papers are edited in a book that is being published by Springer.

Remote sensing development is ultimately driven by applications. Although there were many presentations at the symposium on various applications, because of limited space, only seven papers are published in this special issue. They focus on forest drought monitoring, temporal signature recovery, grass biomass estimation, crop yield estimation, estimation of gross primary productivity related to urbanization, mapping regional evapotranspiration, and land use simulation in a river floodplain.

The first paper addresses the detection of drought conditions of the forest canopy by using hyperspectral remote sensing. Drought has become one of the most serious global problems resulting in water shortages, economic losses, and adverse social consequences. This paper addresses this critical topic and has significant implications for various applications ranging from forest fire hazard monitoring to the assessment of forest productivity. After comparing the spectral signatures of leaves measured in the laboratory and drought-stressed forest canopy from spaceborne Hyperon hyperspectral data, Lee et al. find that the reflectance spectra of leaves do not match the spectral characteristics of the canopy-level and that forest canopy spectra under moderate drought status may be more influenced by forest canopy structure (e.g., canopy closure and leaf area index - LAI) and foliage density, rather than by canopy moisture level.

The second paper presents a new method to recover the temporal signatures from Moderate-Resolution Imaging Spectroradiometer (MODIS) observations. MODIS represents the medium-resolution (about 1km) satellite sensors which usually observe land surfaces multiple times daily. The observations are highly contaminated by the atmosphere, such as clouds and aerosols, which are considered as large sources of noise for land applications. Lu et al. propose a new method to recover the underlying temporal signatures from MODIS observations. Their algorithm is based on the wavelet transformation. They compare their new algorithm with three other popular algorithms at both pixel and image levels and find that this new algorithm outperforms all others. They also apply this new method to MODIS LAI, albedo and normalized difference vegetation index (NDVI) products.

The third paper addresses the estimation of grassland biomass from hyperspectral remote sensing data. Biomass is the preferred measure when assessing plant species abundance, species richness and species evenness. Estimation of above-ground biomass is necessary for studying productivity, carbon cycles, nutrient allocation, and fuel accumulation in terrestrial ecosystems. Remote sensing is the only means to estimate biomass regionally. Clevers et al. estimate grassland biomass from the field measured hyperspectral reflectance spectra using the support vector machine algorithm. Different vegetation indices are used as the predictors. Their results indicate that the vegetation index calculated from one near-infrared band and one band near the red edge produces the best results, much better than any other combinations using the shorter wavebands.

The fourth paper demonstrates the application of remote sensing to crop yield estimation. Crop yield is a key element for rural development and an indicator for national food security. Accurate and objective estimation of crop yield over large areas is critical for national food security through policy making on import/export plans and prices. Physical methods for estimating crop yield using biophysical crop growth models and remote sensing data have been explored, but the statistical methods are currently more reliable. Li et al. present the methodologies for estimating crop yield on the county basis from AVHRR data of more than 20+ years. They explore two methods based on multivariate regression analysis and artificial neural network technique with a new optimization procedure. They can achieve an accuracy as high as 85%.

The fifth paper assesses the impact of urbanization on gross primary productivity (GPP) around southeastern Michigan in the 1990s. Zhao et al. calculate the annual GPP using a radiation efficiency model with some parameters estimated from AVHRR data and ancillary information. They detect the land cover change using Thematic Mapper (TM) and the Enhanced Thematic Mapper Plus (ETM+) imagery. They find that the regional annual GPP increased in southeastern Michigan and believe that it is due to the increased fraction of tree cover throughout the entire region.

The sixth paper examines the use of mapping regional evapotranspiration (ET) from EM/ETM+ imagery. ET plays an important role in energy exchange and hydrological cycle between the land surface and atmosphere. Accurate ET product has wide application in many disciplines, such as geography, meteorology, hydrology and ecology. It is also required by short-term numerical weather predication models and longer-term simulations for climate predication. The conventional point measurements cannot account for the surface heterogeneity. There are two major approaches for mapping ET from remote sensing: 1) estimate from empirical indices or simple equations; and 2) calculate it from the residual of the energy balance equation by first determining other components. Liu et al. apply the second method to estimate regional ET around Beijing City using TM/ETM+ in conjunction with meteorological observations for several years. The validation results indicate they have achieved high accuracy.

The last paper addresses the land use in the river floodplain. Schaepman et al. combine hyperspectral remote sensing derived products with a dynamic vegetation model to improve the simulation and evaluation of future scenarios for a river floodplain. They derive LAI, plant functional type (PFT) distribution, and model dominant species abundances as input and use a dynamic vegetation model to simulate vegetation succession under scenarios of changing abiotic conditions and management regimes. The simulation results agree well with the actual vegetation succession in the area.

Acknowledgements
We would like to thank the International Scientific Committee and the local Organizing Committee for organizing this highly successful symposium and the sponsors for their great support. The reviewers’ valuable comments have greatly improved the presentations of the papers. This special issue is supported in part by a NASA grant under NNG06GG24G.

Guest Editors
Shunlin Liang
Professor
Department of Geography
University of Maryland, USA

Jiyuan Liu
Professor & Director
Institute for Geographical Sciences and Natural Resources Research,
Beijing, China

Xiaowen Li
Professor & Director
Center for Remote Sensing and GIS
Beijing Normal University, China