The effect of changing environmental conditions on microwave signatures of forest ecosystems: preliminary results of the March 1988 Alaskan aircraft SAR experiment

Abstract

In preparation for the first European Space Agency (ESA) Remote Sensing(ERS-I) mission,a series of multitemporal, multifrequency, multipolarization aircraft synthetic aperture radar (SAR) data sets were acquired over the Bonanza Creek Experimental Forest near Fairbanks, Alaska in March, 1988. P-, L- and C-band data were acquired with the NASA/JPL Airborne SAR on five differentdays over a period of two weeks. The airborne data were augmented with intensiveground calibration data as well as detailed, simultaneous in situ measurements of the geometric, dielectric and moisture properties of the snow and forest canopy. During the time period over which the SAR data were collected, the environmental conditions changed significantly; temperatures ranged from unseasonably warm (I to 9°C) to well below freezing (-8 to - 15°C), and the moisture content of the snow and trees changed from a liquid to a frozenstate. The SAR data clearly indicate the radar return is sensitive to these changing environmental factors and preliminary analysis of the L-band SAR data shows a 0·4 to 5·8dB increase (depending on polarization and canopy type) in the radar cross section of the forest stands under the warm conditions relative to the cold. These SAR observations are consistent with predictions from a theoretical scattering model. These preliminary results are presented to illustrate the opportunity afforded by the ERS-l SAR to monitor temporal changes in forest ecosystems.     

Interpretation and model analysis of MAESTRO 1 Flevoland data

A model, based on the Radiative Transfer Theory and the Matrix Doubling algorithm, is tested with agricultural and forest radar returns measured by the JPL AIRSAR sensor over the Flevoland site (The Netherlands) at P, L and C bands.

The test is generally valid, although some discrepancies are noted in the absolute P band values of μ° and not all the input parameters were obtainable from ground measurements. The model is also used to interpret the experimental results and to estimate the relative importance of the different scattering contributions.     

Assessment of stand‐wise stem volume retrieval in boreal forest from JERS‐1 L‐band SAR backscatter

JERS‐1 L‐band SAR backscatter from test sites in Sweden, Finland and Siberia has been investigated to determine the accuracy level achievable in the boreal zone for stand‐wise forest stem volume retrieval using a model‐based approach. The extensive ground‐data and SAR imagery datasets available allowed analysis of the backscatter temporal dynamics. In dense forests the backscatter primarily depended on the frozen/unfrozen state of the canopy, showing a ～4 dB difference. In sparse forests, the backscatter depended primarily on the dielectric properties of the forest floor, showing smaller differences throughout the year. Backscatter modelling as a function of stem volume was carried out by means of a simple L‐band Water Cloud related scattering model. At each test site, the model fitted the measurements used for training irrespective of the weather conditions. Of the three a priori unknown model parameters, the forest transmissivity coefficient was most affected by seasonal conditions and test site specific features (stand structure, forest management, etc.). Several factors determined the coefficient's estimate, namely weather conditions at acquisition, structural heterogeneities of the forest stands within a test site, forest management practice and ground data accuracy. Stem volume retrieval was strongly influenced by these factors. It performed best under unfrozen conditions and results were temporally consistent. Multi‐temporal combination of single‐image estimates eliminated outliers and slightly decreased the estimation error. Retrieved and measured stem volumes were in good agreement up to maximum levels in Sweden and Finland. For the intensively managed test site in Sweden a 25% relative rms error was obtained. Higher errors were achieved in the larger and more heterogeneous forest test sites in Siberia. Hence, L‐band backscatter can be considered a good candidate for stand‐wise stem volume retrieval in boreal forest, although the forest site conditions play a fundamental role for the final accuracy.     

Monitoring snowmelt across the Arctic forest–tundra ecotone using Synthetic Aperture Radar

Detailed snowpack observations, meteorology, topography and landcover classification were integrated with multi‐temporal SAR data to assess its capability for landscape scale snowmelt mapping at the forest–tundra ecotone. At three sites along an approximately 8° latitudinal gradient in the Fennoscandian mountain range, 16 multi‐temporal spaceborne ERS‐2 synthetic aperture radar (SAR) were used for mapping snowmelt.

Comparison of field measurements and backscatter values demonstrates the difficulty of interpreting observed backscatter response because of complex changes in snow properties on diurnal and seasonal temporal scales. Diurnal and seasonal melt–freeze effects in the snowpack, relative to the timing of ERS‐2 SAR image acquisition, effectively reduce the temporal resolution of such data for snow mapping, even at high latitudes.

The integration of diverse data sources did reveal significant associations between vegetation, topography and snowmelt. Several problems with the application of thresholding for the automatic identification of snowmelt were encountered. These largely related to changes in backscattering from vegetation in the late stages of snowmelt. Due to the impact of environmental heterogeneity in vegetation at the forest–tundra ecotone, we suggest that the potential to map snow cover using single polarization C‐band SAR at the forest–tundra ecotone may be limited to tundra areas.     

Sensitivity of space-borne SAR data to forest parameters over sloping terrain. Theory and experiment

Recently, SAR data proved to be useful for the retrieval of forest biomass. However, the effects of terrain slope must be addressed towards the generalization of biomass retrieval for varied forest and environmental conditions. To this aim, we developed experimental and theoretical approaches allowing the study of multi-frequency/multi-polarization forest backscatter of a given forest type, as a function of forest parameters and SAR local incidence angle over the relief. The experimental results showed that the sensitivity of SAR data to biomass was similar to that obtained over a flat terrain, only if the backscatter data were calibrated for slope effects. Moreover, the backscatter must also be corrected for its angular decrease, which can be removed using a simple angular model developed under assumptions of theoretical equations. The highest correlation of corrected backscatter with forest parameters related to aboveground biomass (such as stand age and bole volume) was achieved at L-HV 55° (R ²     

Assessment of the ASAR sensor radiometric quality in comparison to ERS‐2 and RADARSAT‐1 SAR data

The estimation of geophysical parameters from Synthetic Aperture Radar (SAR) data necessitates well‐calibrated sensors with good radiometric precision. In this paper, the radiometric calibration of the new Advanced Synthetic Aperture Radar (ASAR‐ENVISAT) sensor was assessed by comparing ASAR data with ERS‐2 and RADARSAT‐1 SAR data. By analysing the difference between radar signals of forest stands, the results show differences of varying importance between the ASAR on the one hand, and the ERS‐2 and the RADARSAT‐1 on the other. For recent data acquired at the end of 2005, the difference varies from ?0.72 to +0.72 dB, with temporal variations that can reach 1.1 dB. For older data acquired in 2003 and 2004, we observe a sharp decrease in the radar signal in the range direction, which can attain 3.5 dB. The use of revised calibration constants provided recently by the European Space Agency (ESA) significantly improves the results of the radiometric calibration, where the difference between the ASAR and the other SARs will be lower than 0.5 dB.     

Demonstration of an improved calibration scheme for ERS-l SAR imagery using a scatterometer wind model

The importance of an improved calibration scheme for the derivation of the normalized radar cross-section coefficient σ° for distributed targets using ERS-l SAR imagery is assessed. The improved calibration scheme includes corrections relating to the saturation of the on-board Analogue to Digital Convertor (ADC) and to variations in the replica pulse power. The significance and effectiveness of the additional corrections is demonstrated by comparing the range variation of σ° from 10 ERS-l.SAR.PRI images of the English Channel with that predicted by the CMOD4 scatterometer wind model. Close agreement is found at all range positions, provided the additional corrections are applied.     

Retrieval of growing stock volume in boreal forest using hyper-temporal series of Envisat ASAR ScanSAR backscatter measurements

Methods for the estimation of forest growing stock volume (GSV) are a major topic of investigation in the remote sensing community. The boreal zone contains almost 30% of global forest by area but measurements of forest resources are often outdated. Although past and current spaceborne synthetic aperture radar (SAR) backscatter data are not optimal for forest-related studies, a multi-temporal combination of individual GSV estimates can improve the retrieval as compared to the single-image case. This feature has been included in a novel GSV retrieval approach, hereafter referred to as the BIOMASAR algorithm. One innovative aspect of the algorithm is its independence from in situ measurements for model training. Model parameter estimates are obtained from central tendency statistics of the backscatter measurements for unvegetated and dense forest areas, which can be selected by means of a continuous tree canopy cover product, such as the MODIS Vegetation Continuous Fields product. In this paper, the performance of the algorithm has been evaluated using hyper-temporal series of C-band Envisat Advanced SAR (ASAR) images acquired in ScanSAR mode at 100 m and 1 km pixel size. To assess the robustness of the retrieval approach, study areas in Central Siberia (Russia), Sweden and Québec (Canada) have been considered. The algorithm validation activities demonstrated that the automatic approach implemented in the BIOMASAR algorithm performed similarly to traditional approaches based on in situ data. The retrieved GSV showed no saturation up to 300 m³/ha, which represented almost the entire range of GSV at the study areas. The relative root mean square error (RMSE) was between 34.2% and 48.1% at 1 km pixel size. Larger errors were obtained at 100 m because of local errors in the reference datasets. Averaging GSV estimates over neighboring pixels improved the retrieval statistics substantially. For an aggregation factor of 10 × 10 pixels, the relative RMSE was below 25%, regardless of the original resolution of the SAR data.     

Fully polarimetric airborne SAR and ERS SAR observations of snow: implications for selection of ENVISAT ASAR modes

Snow cover has a substantial impact on processes involved in the interaction between atmosphere and surface, and the knowledge of snow parameters is important in both climatology and weather forecasting. With the upcoming launch of Advanced Synthetic Aperture Radar (ASAR) instruments on Envisat, enhanced snow-mapping capabilities are foreseen. In this paper fully polarimetric C- and L-band airborne SAR data, ERS SAR and auxiliary data from various snow conditions in mountainous areas are analysed in order to determine the optimum ASAR modes for snow monitoring. The data used in this study are from the Norwegian part of the snow and ice experiment within the European Multi-sensor Airborne Campaign (EMAC'95) acquired in the Kongsfjellet area, located in Norway, 66°?N, 14°?E. Fully polarimetric C- and L-band SAR data from ElectroMagnetic Institute SAR (EMISAR), an airborne instrument operated by the Danish Center for Remote Sensing (DCR), were acquired in March, May, and July 1995. In addition, several ERS SAR, airborne photos, field and auxiliary data were acquired.

A larger separation between wet snow and bare ground in EMISAR C-VV polarisation data was found at high incidence angle (55°) compared to lower incidence angle (45°). Cross-polarized observations from bare ground, dry and wet snow in the incidence angle range 35° to 65° are below the specified Envisat ASAR noise floor of –20–22 dB. The backscattering angular dependency for wet snow and bare ground derived from EMISAR C-VV and ERS SAR data corresponds well, and agrees to some extent with volume and surface scattering model results. The C-band is more sensitive to variation in snow properties than the L-band.     

简述极化SAR定标处理技术研究进展

极化定标是极化SAR数据用于图像解译和定量参数反演的关键步骤。综述了国内外极化SAR定标技术及应用研究的主要成果,着重介绍极化SAR定标技术发展以来取得的多种关键算法进展,系统梳理包括点目标定标算法、分布目标定标算法、法拉第旋转校正算法等技术原理介绍、参数定义估计及技术脉略发展分析,并讨论各技术方法在机载及星载SAR系统定标领域取得的应用发展现状,同时逐步分解该技术领域目前面临的技术难点及算法发展。最后探讨分析极化SAR定标技术发展中的未来需求难点和继续需要解决的问题,为研究人员进一步推动极化SAR定标技术发展提供参考。     

Vicarious calibration of GOCI for the SeaDAS ocean color retrieval

A key on-orbit calibration step for satellite remote sensing of ocean color is the vicarious calibration. This establishes the final gains for each spectral band on the sensor that minimize bias in the retrieved ocean color signal. The vicarious calibration is specific to the instrument and the atmospheric correction algorithm. The vicarious calibration gains for the Geostationary Ocean Color Imager (GOCI) are presented here, which were derived to optimize the performance of NASA’s standard atmospheric correction algorithm as implemented in the l2gen code and distributed through the SeaDAS open-source software package. Following NASA’s protocols, the near-infrared (NIR) bands were calibrated first, and the visible bands were then calibrated relative to this fixed NIR calibration. The gain for the 745-nm NIR band was derived using a fixed aerosol model, which was chosen based on the Angstrom Coefficients derived from MODIS on Aqua (MODISA). For the vicarious gains of the visible bands, two sources for the target water-leaving radiances were tested: matchups from MODISA and climatological data from SeaWiFS. A validation analysis using AERONET-OC data shows an improvement in sensor performance when compared with results using the current vicarious gains and results using no vicarious calibration. Good agreement was found in vicarious gains derived using both concurrent MODISA and climatological SeaWiFS as vicarious calibration data sources. These results support the use of a concurrent sensor for the vicarious calibration when in situ data are not available and demonstrate that using climatology from a well-calibrated sensor like SeaWiFS for the vicarious calibration is a valid alternative when it is not possible to use a concurrent sensor or in situ data. We recommend using the gains derived from concurrent GOCI matchups with MODISA for GOCI processing in SeaDAS/l2gen.     

水面船只SAR探测的极化方式研究

利用SIR-C SAR的C和L波段全极化数据,分析水面船只的极化散射特性和船只与背景海面雷达后向散射的信噪比特性,研究水面船只SAR探测的最优极化方式。结果显示,二面角散射是水面船只SAR成像的主要机理。线性极化中,HV极化具有最大的船只与背景海面雷达后向散射信噪比。与线性极化相比,圆极化的雷达后向散射信噪比更优。C波段和L波段的水面船只的极化散射特性存在较大的差异,L波段的信噪比大于C波段的信噪比。水面船只的雷达后向散射特性表明,L波段的圆极化是水面船只探测的最优极化方式。     

Potential of SAR for forest bole volume estimation

MAESTRO I data from the Flevoland lest site in the Netherlands have been used for this study. From a complementary ground data collection, the bole volumes of a large number of stands of mainly poplar and ash have been estimated. The relationships between radar backscattering and bole volumes have been examined experimentally and theoretically. In the case studied, the radar backscattering sensitivity to bole volumes increases as the wavelengths increase. and is highest at P band. The sensitivity of the radar backscattering to variations of forest canopy components and moisture contents has been investigated theoretically at P band. It is important to obtain information on such variations before the inversion problem can be solved. The present study indicates a potential for bole volume determination by P-band SAR.     

Power spectrum analysis of SAR data for spatial forest characterization in Amazonia

Power spectrum analysis was used for the analysis of spatial forest features from airborne X‐band synthetic aperture radar (SAR) data in the Brazilian Amazon. Spectral estimates were arrived at empirically by periodograms and correlograms, and from autoregressive moving‐average (ARMA) models. The spectral estimates derived from SAR data were validated by those derived from ground data with locational match. The results obtained by ARMA modelling revealed particularly good correspondence between remote sensing and reference data: repeating patterns at pixel level could be detected in the images. These patterns were shown to arise from canopy structure and distances between major tree individuals; and thus allowed the extraction of parameters of spatial forest structure, particularly of forest density. The method was applied to an example area of primary tropical forest, and its spatial patterns were modelled.     

Assessing relationships between Radarsat-2 C-band and structural parameters of a degraded mangrove forest

Relationships were assessed between mangrove structural data (leaf area index (LAI), stem density, basal area, diameter at breast height (DBH)) collected from 61 stands located in a black mangrove (Avicennia germinans)-dominated forest and both single polarized ultra-fine (3 m) and multipolarized fine beam (8 m) Radarsat-2 C-band synthetic aperture radar (SAR) data. The stands examined included representatives from the four types of mangroves that typify this degraded system, specifically: predominantly dead mangrove, poor-condition mangrove, healthy dwarf mangrove, and tall healthy mangrove. The results indicate that the selection of the spatial resolution (3 m vs. 8 m) of the incidence angle (27–39°) and the polarimetric mode greatly influence the relationship between the SAR and mangrove structural data. Moreover, the extent of degradation, i.e. whether dead stands are considered, also determines the strength of the relationships between the various SAR and mangrove parameters.

When dead stands are included, the strongest overall relationships between the ultra-fine backscatter (incidence angle of ～32°) and the various structural parameters were found using the horizontal-horizontal (HH) polarization/horizontal-vertical (HV) polarization ratio. However, if the dead stands are not included, then significant relationships with the ultra-fine data were only calculated with the HH data. Similar results were observed using the corresponding incidence angle (～33°) of the fine beam data. When a shallower incidence angle was considered (～39°), fewer and weaker relationships were calculated. Moreover, no significant relationships were observed if the dead stands were excluded from the sample at this incidence angle. The highest correlation coefficients using the steepest incidence (～27°) were found with the co-polarized (HH, vertical-vertical (VV) polarization) modes. Several polarimetric parameters (entropy, pedestal height, surface roughness, alpha angle) based on the decomposition of the scattering matrix of the fine beam mode at this incidence angle were also found to be significantly correlated to mangrove structural data. The highest correlation (R = 0.71) was recorded for entropy and LAI. When the dead stands were excluded, volume scattering was found to be the most significant polarimetric parameter. Finally, multiple regression models, based on texture measures derived from both the grey level co-occurrence matrix (GLCM) and the sum and difference histogram (SADH) of the ultra-fine data, were developed to estimate mangrove parameters. The results indicate that only models derived from the HH data are significant and that several of these were strong predictors of all but stem density.     

Radar-based forest biomass estimation

The potential of radar-based tree biomass estimation has been studied using polarimetric SAR data from the Freiburg test site of the MAESTRO I Campaign and scatterometer data from a test site in Finland. Using the Freiburg SAR data and polarizaiton synthesis, the most suitable polarization combination was obtained. In P band, the maximum correIalions, which were found ncar the linear H V polarization, were up to 0·75.

In the Finnish test site, a strong negative correlation “correlation coefficient -0·65” existed between the pine biomass and X-VV backscatter. When the combination of X and C bands “measured by the HUTSCAT seatterometer” was used, a correlation coefficient of 0·81 was obtained.     

一种全实时SAR方位向数字预处理滤波器的设计

利用机载SAR方位向预处理的基本原理，巧妙地将方位向FIR滤波和降采样相结合，从而提出了一种SAR方位向预处理实时快速算法的VLSI结构，并利用HDL语言将该结构综合在了一片大规模CPLD上，得以硬件实现，外围辅助其它单元电路，研制成功了面向地学高分辨率机载SARY方位向预处理分机。经过使用证明，该分机可以适用于多种波段的机载SAR系统，达到了满意的处理效果。     

Classification of forest volume resources using ERS tandem coherence and JERS backscatter data

Considering recent progress in the development of techniques and methods to achieve biomass estimates and full carbon accounting, remote sensing research of forested ecosystems needs to be aimed towards the retrieval of information at global scales. In this paper, an algorithm for the estimation of growing stock volume, an important parameter for the commercial forest community and a proxy for woody biomass density, from ERS and JERS synthetic aperture radar (SAR) data is described. The algorithm is based on the information content of both ERS tandem coherence and JERS backscatter images and was developed using ground data, made available by the Russian Forestry Services. It is tested on SAR datasets of boreal forests in Siberia, a managed, temperate forest plantation in the United Kingdom and a semi-natural boreal forest at Siggefora in Sweden. Comparisons of the classified products, comprising three growing stock interval classes and one non-forest class are made with ground data. The results of this accuracy assessment exercise show that the algorithm yields, in all cases, overall classification accuracies of greater than 70%. A visual comparison is made of the algorithm performance over a tropical forest region of Brazil. The results indicate that the algorithm has the potential to retrieve growing stock volume estimates in forest ecosystems throughout the globe.     

Development of a forest canopy height estimation model using GLAS full waveform data over sloping terrain

ABSTRACT

The accurate estimation of forest canopy height is important because it leads to increased accuracy in the estimation of biomass, which is used in the study of the global carbon cycle, forest productivity, and climate change. However, there is no well-developed model that accurately estimates canopy height over undulating land. This paper describes the development of a back-propagation (BP) neural network model that estimates forest canopy height more accurately than other types of model. For modeling purposes, the land in the study area was classified as either plain (low relief areas) or hilly (high relief areas). Four different slope partition thresholds (5°, 10°, 15°, and 20°) were tested to determine the most suitable boundary value. ICESat-GLAS data provided by the Geoscience Laser Altimeter System (GLAS) aboard the Ice, Cloud and Land Elevation Satellite (ICESat), field survey data, and digital elevation model (DEM) data were collected and refined, and various parameters, including waveform extent and topographic index, were calculated. A BP neural network model was created to estimate forest canopy height. Two other models were also developed, one using the topographic index and the other using multiple linear regression, for comparison with the BP neural network model. After calibration, the three models were tested to assess the accuracy of the estimates. The results showed that the BP model estimated canopy height more accurately than the other two models. The use of a 10° boundary to partition the topography into low relief areas and high relief areas improved the accuracy of each model; using the 10° slope boundary, the coefficient of correlation r between the estimates given by the BP neural network model and the field-measured data increased from 0.89 to 0.95 and the Root Mean Square Error (RMSE) decreased from 1.01 to 0.73 m.     

Analysis of different polarimetric target decomposition methods in forest density classification using C band SAR data

The ability of synthetic aperture radar (SAR) C-band microwave energy to penetrate within forest vegetation makes it possible to extract information on crown components, which in turn gives a better approximation of relative canopy density than optical data-derived canopy density. Many studies have been reported to estimate forest biomass from SAR data, but the scope of C-band SAR in characterizing forest canopy density has not been adequately understood with polarimetric techniques. Polarimetric classification is one of the most significant applications of polarimetric SAR in remote sensing. The objective of the present study was to evaluate the feasibility of different polarimetric SAR data decomposition methods in forest canopy density classification using C-band SAR data. Landsat (Land Satellite) 5 TM (Thematic Mapper) data of the same area has been used as optical data to compare the classification result. RADARSAT (Radar Satellite)-2 image with fine quad-pol obtained on 27 October 2011 over tropical dry forests of Madhav National Park, India, was used for the analysis of full polarimetric data. Six decomposition methods were selected based on incoherent decomposition for generating input images for classification, i.e. Huynen, Freeman and Durden, Yamaguchi, Cloude, Van zyl, and H/A/α. The performance of each decomposition output in relation to each land cover unit present in the study area was assessed using a support vector machine (SVM) classifier. Results show that Yamaguchi 4-component decomposition (overall accuracy 87.66% and kappa coefficient (κ) 0.86) gives better classification results, followed by Van Zyl decomposition (overall accuracy 87.20% and κ 0.85) and Freeman and Durden (overall accuracy 86.79% and κ 0.85) in forest canopy density classification. Both model-based decompositions (Freeman and Durden and Yamaguchi4) registered good classification accuracy. In eigenvector or eigenvalue decompositions, Van zyl registered the second highest accuracy among different decompositions. The experimental results obtained with polarimetric C-band SAR data over a tropical dry deciduous forest area imply that SAR data have significant potential for estimating canopy density in operational forestry. A better forest density classification result can be achieved within the forest mask (without other land cover classes). The limitations associated with optical data such as non-availability of cloud-free data and misclassification because of gregarious occurrence of bushy vegetation such as Lantana can be overcome by using C-band SAR data.