No full-text available
To read the full-text of this research,
you can request a copy directly from the authors.
Potential of SAR for forest bole volume estimation
International Journal of Remote Sensing
Abstract
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.
... On the second step, the model inversion for the -th stand with backscattering coefficient , and stand-wise stem volume estimation for the total test site area is performed as: (5) However, in, e.g., [24], [45] fitting was performed using the forward model in the manner given by (4). Here, we use the inverted model (5) in the fitting procedure, in order to avoid some of the drawbacks, associated with inversion of the model fitted by (4). ...
... On the second step, the model inversion for the -th stand with backscattering coefficient , and stand-wise stem volume estimation for the total test site area is performed as: (5) However, in, e.g., [24], [45] fitting was performed using the forward model in the manner given by (4). Here, we use the inverted model (5) in the fitting procedure, in order to avoid some of the drawbacks, associated with inversion of the model fitted by (4). In particular, this can help with the problem of the impossibility of estimating stem volume for some plots, where , or producing negative stem volume estimates under ...
... where is an optimal estimate of stem volume of forest stand, is a vector of length of stem volume estimates of the same -th stand from individual SAR images according to (5). Parameter is required to compensate for systematic bias, expected due to non-accounted scattering mechanisms in model (1), e.g., the missing contribution from ground-trunk interaction. ...
This paper presents a modified robust stem volume retrieval approach suitable for use with L-band SAR imagery. Multitemporal dual-polarization SAR imagery acquired by ALOS PALSAR during the summer-autumn 2007 is used in the study, along with stand-wise forest inventory data from two boreal forest sites situated in central Finland. The average sizes of forest stands at the study sites were 3 ha and 4.8 ha. The method used employs model fitting with an inverted semi-empirical boreal forest model, and takes advantage of the multitemporal aspect in order to improve the stability and accuracy of stem volume estimation. Multitemporal combination of model output in a multivariate regression framework allows volume estimates to be obtained with an RMSE about 43% of the mean of 110 m3 /ha, and a coefficient of determination R 2 of 0.71 in the best case. The methodology used can be employed to produce large-area stem volume maps from dual-polarization ALOS PALSAR imagery mosaics.
... On the other hand, a strong inverse relationship exists between pine biomass and X-band VV-polarisation in the Finnish test site. In the Flevoland forest region of Netherlands, the AIRSAR data with modelling have shown potential for growing stock volume estimation particularly at longer wavelengths (Israelsson et al. 1994). This site is broad-leaved and mixed species with poplar, ash, oak, maple, willow and beech as the main species. ...
... In (Germany), Flevoland (The Netherlands) and Thetford (U.K.) focusing on SAR bands at P-, L-and C-band frequencies (Baker et al. 1994, Basili et al. 1994, De Grandt et al. 1994, Dobson et al. 1992, Foody et al. 1994, Israelsson et al. 1994, Rauste et al. 1994). , Hajnsek et al. 2009). ...
https://lra.le.ac.uk/handle/2381/37924
... Furthermore, the nature of microwave interaction with forest is sensitive to its forest biomass as a result of its geometry etc., (Baker et al., 1994;Beaudoin et al., 1994;Imhoff, 1995a;Kasischke et al., 1995). According to Le Toan et al., 1992, as most of the backscatter energy originates from the upper crown layer, in Cband, it is of practically not much useful for biomass estimation due to shorter wavelength (e.g., Radarsat, SIR-C, ERS-1), Israelsson, 1994. There is a strong relationship between SAR backscatter L-band around 40-60 t/ha. ...
... Campaign , Israelsson, 1994. RADARSAT provides data at different incidence angles. ...
... Furthermore, the nature of microwave interaction with forest is sensitive to its forest biomass as a result of its geometry etc., (Baker et al., 1994;Beaudoin et al., 1994;Imhoff, 1995a;Kasischke et al., 1995). According to Le Toan et al., 1992, as most of the backscatter energy originates from the upper crown layer, in Cband, it is of practically not much useful for biomass estimation due to shorter wavelength (e.g., Radarsat, SIR-C, ERS-1), Israelsson, 1994. There is a strong relationship between SAR backscatter L-band around 40-60 t/ha. ...
... Campaign , Israelsson, 1994. RADARSAT provides data at different incidence angles. ...
... The benefit of synthetic aperture radar (SAR) data for forestry applications has been demonstrated by numerous studies. While techniques aimed at forest cover monitoring and clear-cut detection are able to deliver valuable information in the boreal zone Yatabe and Leckie, 1995;Rignot et al., 1997;Thiel et al., 2006;Thiel et al., 2009), the estimation of forest growing stock volume (GSV) is still challenging because of problems related to saturation and considerable estimation uncertainties (Dobson et al., 1992;Israelsson et al., 1994;Santoro et al., 2006). In Dobson et al. (1992), Israelsson et al. (1994), and Santoro et al. (2006) forest biomass estimation was based on backscatter intensity. ...
... While techniques aimed at forest cover monitoring and clear-cut detection are able to deliver valuable information in the boreal zone Yatabe and Leckie, 1995;Rignot et al., 1997;Thiel et al., 2006;Thiel et al., 2009), the estimation of forest growing stock volume (GSV) is still challenging because of problems related to saturation and considerable estimation uncertainties (Dobson et al., 1992;Israelsson et al., 1994;Santoro et al., 2006). In Dobson et al. (1992), Israelsson et al. (1994), and Santoro et al. (2006) forest biomass estimation was based on backscatter intensity. In other studies, multitemporal data proved to have the potential to reduce these shortcomings (Koskinen et al., 2001;Santoro et al., 2002;Askne et al., 2003;Eriksson et al., 2003;Pulliainen et al., 2003); in these papers, the magnitude of repeat pass coherence was used as the biomass estimator. ...
This paper investigates the impact of freezing on the properties of the magnitude of the interferometric coherence |γ| in central Siberia and discusses its implications for forest growing stock volume (GSV) estimation in the boreal zone. Eighty-seven acquisitions were employed and approximately 300 interferograms were generated. Accordingly, a high statistical credibility of the obtained results can be presumed. The temporal baselines of the interferograms ranged from 46 days to 2. 5 years. The random volume over ground model was applied to support the interpretation of the observations. Compared with unfrozen conditions, we observed increased coherence over open areas, decreased coherence over dense forest, decreased spread of coherence, and improved correlation between |γ| and GSV during a frozen state. At frozen conditions, the experimental data showed no proof that the perpendicular baseline B impacted |γ| over dense forest, whereas at unfrozen conditions an impact was detected. Consequently, at frozen conditions, the temporal decorrelation was the major source of decorrelation and obstructed the detection of volume decorrelation effects. Nevertheless, |γ| acquired at frozen state exhibits potential for GSV mapping. The relationship between GSV and |γ| can be described with an average coefficient of determination R2 of 0. 6. Saturation occurs at about 250 m3/ha.
... The forest cover change map of Thuringia was created using the post-classifi cation approach. Five JERS SAR images, from 1992, 1993, 1994, 1995and 1998, were selected. The forest cover maps were formed by thresholding σ 0 for each individual SAR image, and the classifi cation for 1992 is shown as Figure 7 . ...
... This results in reduced accuracy in mapping reforestation. Polarimetric L-band SAR data (or at least, other polarizations in addition to HH) are expected to allow higher accuracy in this application ( Israelsson et al. , 1994 ;Rignot et al. , 1997 ). ...
... 2001, forest disturbance mapping (e.g. logging, forest fire, and wind damage) , Yatabe & Leckie 1995, Rignot et al. 1997) and forest biomass assessment (Dobson et al. 1992, Israelsson et al. 1994, Santoro et al. 2006. Lower radar frequencies proved to be of particular adequacy. ...
... As their backscatter lies close to the threshold that separates forest from non-forest, these stands cannot be clearly assigned to either class (forest/non-forest). The usage of polarimetric L-band SAR data is expected to remedy this deficiency (Israelsson et al. 1994, Rignot et al. 1997. Hence, the effect of this data type has to be investigated in further studies. ...
This paper assesses the feasibility of forest cover mapping and the delineation of deforestation using Japanese Earth Resource Satellite (JERS-1) Synthetic Aperture Radar (SAR) data. The assessment is carried out at five test sites in Germany (Thuringia), the UK (Kielder), Sweden (Remningstorp and Brattåker) and Russia (Chunsky). These temperate and boreal sites all have high forest cover, but with different forestry management practices. The stands at the Swedish, Russian and UK sites are harvested by clearcutting, while in Thuringia, thinning is the predominant practice. Man-made deforestation is characterised in SAR imagery by regular geometric patterns which can be segmented and classified for data analysis. This reduces the statistical effects of SAR speckle. The procedure for mapping deforested areas exploits time series of SAR images, taken from the period 1992-1998 during which JERS was operational. Two different approaches were developed. The first detects forest cover separately for each JERS scene, while the second takes all scenes into account simultaneously. Images are classified into forest, non-forest and deforested areas. The overall accuracy of the derived forest cover map is about 90 % in acreage, and about 90 % for logging. Two different approaches to detect forest cover changes have been applied. The post-classification detection of changes in forest cover is based on the analysis of the delineated forest cover maps. The forest cover maps are derived for each chosen JERS scene. A temporal change of the classified forest cover can be interpreted as ARD activity. Knowledge based rules were used for this analysis. The pre-classification detection of changes in forest cover utilises all chosen JERS images at once. By means of a multitemporal
... 2003). Desde a década de 90 a quantificação de determinadas variáveis biofísicas da cobertura florestal vem sendo realizada com abordagens que empregam dados de radar de abertura sintética polarimétrica. No que se refere à estimativa volumétrica, alguns estudos já foram conduzidos em plantios de coníferas e florestas boreais, segundo Israelsson et al. (1994). De acordo com Henderson e Lewis (1998), os estudos de variáveis biofísicas da cobertura florestal são frequentemente baseados apenas nas medidas de potência do radar, adquiridas nas três polarizações lineares (HH, VV e HV). Variáveis biofísicas são importantes descritores de florestas nativas ou implantadas. Por exemplo, altura e diâme ...
... rização paralela (Rp) e a razão de polarização cruzada (Rc), descritos em Henderson e Lewis (1998); a potência total (PT), apresentada por Boerner et al. (1991); e os índices de biomassa (BMI, biomass index), de estrutura do dossel (CSI, canopy structure index) e de espalhamento volumétrico (VSI, volume scattering index), desenvolvidos por Pope et al. (1994) ...
The availability of orbital data has brought new perspectives to both academic and corporative sectors solving demands in forestry science. The main objective of this work was to evaluate the potential of radar data for estimating the timber volume in Eucalyptus spp stands. For this, the correlations between spectral data and timber volume were performed following the use of Multiple Linear regression models as the descriptors of the relationships established. Images were obtained from radar sensors PALSAR (Phased Array type L-band Synthetic Aperture Radar) aboard the ALOS satellite in areas of eucalyptus production located in Vale do Rio Doce -State of Minas Gerais, Brazil, belonging to CENIBRA. Trees between 4 and 8 years old were selected for the study. The results pointed out that the volume from the stands ware correlated with the biomass index –BMI (-65%) and total power –Pt (-69). The final results pointed out that the volume variable was better explanied by the regression with the combination of Pt and BMI index. The correlation coefficient showed a median value of 0.49. The results achieved in this study support the hypothesis that data coming from PALSAR, sensors can be used to quantify volume variable in eucalyptus plantations. Palavras-chave: sensor palsar, sensoriamento remoto, variáveis biofísicas, palsar sensor, remote sensing, biophysical variables.
... 2001, forest disturbance mapping (e.g. logging, forest fire, and wind damage) , Yatabe & Leckie 1995, Rignot et al. 1997) and forest biomass assessment (Dobson et al. 1992, Israelsson et al. 1994, Santoro et al. 2006. Lower radar frequencies proved to be of particular adequacy. ...
... As their backscatter lies close to the threshold that separates forest from non-forest, these stands cannot be clearly assigned to either class (forest/non-forest). The usage of polarimetric L-band SAR data is expected to remedy this deficiency (Israelsson et al. 1994, Rignot et al. 1997. Hence, the effect of this data type has to be investigated in further studies. ...
In this chapter forest cover and forest cover change mapping basing on image objects is discussed. Change relates to recent complete harvesting and reestablishment, degradation or thinning is not considered. For the change maps two different strategies are proposed. The first one derives the changes by means of previously classified images of a multitemporal dataset and is thus referred to as “post-classification change detection”. For increasing the accuracy of the change maps a knowledge based change detection approach is introduced. The second strategy considers all scenes of the multitemporal dataset simultaneously. This method is referred to as “multidate classification”.
Generally any kind of Earth Observation (EO) data allowing the grey value based separation of forest and non-forest can be applied with both strategies. In this study, JERS-1 (Japanese Earth Resource Satellite) SAR data are used for method development. The feasibility assessment of both object based mapping strategies is performed at five test sites: Germany (Thuringia), UK (Kielder), Sweden (Remningstorp and Brattåker) and Russia (Chunsky). Due to the specific data requirements (broad multitemporal dataset) the first approach could only be successfully implemented at the Thuringia site. It was also tested at Kielder, but with deficient results.
The chapter concludes with the successful realisation of the approach at the Russian service case of GSE FM. Because of the given time frame (1990-recent) other EO data sources had to be implemented. As historical EO data source LANDSAT TM was selected, recent information is derived from ASAR APP.
... In multi-frequency investigations, substantially better results have been obtained at L-and P-band. The sensitivity of the L-band backscatter to forest attributes was higher than at C-band but still lower than at P-band, which made several authors conclude that P-band is most suitable for the estimation of forest parameters (DOBSON et al., 1992; LE TOAN et al., 1992; BEAUDOIN et al., 1994; ISRAELSSON et al., 1994; RANSON & SUN, 1994; RAUSTE et al., 1994; RIGNOT et al., 1994; DOBSON et al., 1995; IMHOFF, 1995; RANSON et al., 1997). An even stronger correlation to forest parameters has been found at VHF and virtually no saturation of the signal has been measured. ...
... Compared to C-band, the higher sensitivity of L-band backscatter to forest attributes is clearly shown by the higher saturation level and correlation observed in several experiments, not only in boreal forests (DOBSON, et al., 1992; LE TOAN et al., 1992; ISRAELSSON et al., 1994; RIGNOT et al., 1994; IMHOFF, 1995; HALLIKAINEN, et al., 1998). Using EMISAR data at two test sites located in Britain and in Sweden close to Kättböle, (BAKER & LUCKMAN, 1999) reported a higher dynamic range at L-band, although the Swedish test site showed saturation at lower stem volumes probably because of the rougher and wetter conditions of the ground. ...
The thesis describes investigations concerning the evaluation of ERS and JERS SAR images and repeat-pass interferometric SAR images for the retrieval of biophysical parameters in boreal forests. The availability of extensive data sets of images over several test sites located in Sweden, Finland and Siberia has allowed analysis of temporal dynamics of ERS and JERS backscatter and coherence, and of ERS interferometric phase. Modelling of backscatter, coherence and InSAR phase has been performed by means of the Water Cloud Model (WCM) and the Interferometric Water Cloud Model (IWCM); sensitivity analysis and implications for the retrieval of forest biophysical parameters have been thoroughly discussed. Model inversion has been carried out for stem volume retrieval using ERS coherence, ERS backscatter and JERS backscatter, whereas for tree height estimation the ERS interferometric phase has been used. Multi-temporal combination of ERS coherence images, and to a lesser extent of JERS backscatter images, can provide stem volume estimates comparable to stand-wise ground-based measurements. Since the information content of the interferometric phase is strongly degraded by phase noise and uncorrected atmospheric artefacts, the retrieved tree height shows large errors.
... In multi-frequency investigations, substantially better results have been obtained at L-and P-band. The sensitivity of the L-band backscatter to forest attributes was higher than at C-band but still lower than at P-band, which made several authors conclude that P-band is most suitable for the estimation of forest parameters (DOBSON et al., 1992;LE TOAN et al., 1992;BEAUDOIN et al., 1994;ISRAELSSON et al., 1994;RANSON & SUN, 1994;RAUSTE et al., 1994;RIGNOT et al., 1994;DOBSON et al., 1995;IMHOFF, 1995;RANSON et al., 1997). An even stronger correlation to forest parameters has been found at VHF and virtually no saturation of the signal has been measured. ...
... Compared to C-band, the higher sensitivity of L-band backscatter to forest attributes is clearly shown by the higher saturation level and correlation observed in several experiments, not only in boreal forests (DOBSON, et al., 1992;LE TOAN et al., 1992;ISRAELSSON et al., 1994;RIGNOT et al., 1994;IMHOFF, 1995;HALLIKAINEN, et al., 1998). Using EMISAR data at two test sites located in Britain and in Sweden close to Kättböle, (BAKER & LUCKMAN, 1999) reported a higher dynamic range at L-band, although the Swedish test site showed saturation at lower stem volumes probably because of the rougher and wetter conditions of the ground. ...
... This relatively low frequency, compared to existing spaceborne SARs, allows the emitted electromagnetic waves to penetrate the canopy and reflect off larger structures such as branches and tree stems where the majority of a tree's biomass is located. This makes P-band especially sensitive to above-ground forest biomass [7]- [9]. A consequence of the increased canopy penetration is that the ground contributes significantly to the total backscattered field, either through direct rough surface scattering or double-bounce scattering by the ground and tree trunks [10]. ...
Temporal variations in synthetic aperture radar (SAR) backscatter over forests are of concern for any SAR mission with the goal of estimating forest parameters from SAR data. In this article, a densely sampled, two-year long time series of P-band (420 to 450 MHz) boreal forest backscatter, acquired by a tower-based radar, is analyzed. The experiment setup provides time series data at multiple polarizations. Tomographic capabilities allow the separation of backscatter at different heights within the forest. Temporal variations of these multi-polarized, tomographic radar observations are characterized and quantified. The mechanisms studied are seasonal variations, effects of freezing conditions, diurnal variations, effects of wind and the effects of rainfall on backscatter. An emphasis is placed on upper-canopy backscatter, which has been shown to be a robust proxy for forest biomass. The canopy backscatter was more stable than ground-level backscatter during non-frozen conditions, supporting forest parameter retrieval approaches based on tomography or interferometric ground notching. Large backscatter variations during frozen conditions, which may be detected using cross-polarised backscatter observations, can result in large errors in forest parameter estimates. Diurnal backscatter variations observed during hot periods were likely connected to tree water transport and storage mechanisms. Backscatter changes were also observed during strong winds. These variations were small in comparison to the variations due to freeze-thaw and soil moisture changes and should not result in significant forest parameter estimation errors. The presented results are useful for designing physically based and semi-empirical scattering models that account for temporal changes in scattering characteristics.
... For the last three decades, most of the research related to using remote-sensing technology as a complementary tool for forest inventories has been dedicated to estimating forest stand parameters using correlation studies (Lu et al. 2004;Alrababah and Alhamad 2011). This is mainly based on deriving the link between the remotely sensed spectral response of a particular ecosystem and its basic features such as basal area, canopy cover, DBH, mean tree height, biomass, and leaf area index (Houborg et al. 2015;Zhao et al. 2016), using multispectral optical satellites datasets (Yu, Shi, and Gong 2015;Zhang, Chen, and Lu 2015;Chen et al. 2015), or through analysing synthetic aperture radar imagery (Israelsson, Askne, and Sylvander 1994;Fransson 1999;Castel et al. 2002;Lehmann et al. 2015;Ningthoujam et al. 2016;Abdullahi, Kugler, and Pretzsch 2016). ...
Forest parameters, such as mean diameter at breast height (DBH), mean stand height (H) or volume per hectare (V), are imperative for forest resources assessment. Traditional forest inventory that is usually based on fieldwork is often difficult, time-consuming, and expensive to conduct over large areas. Therefore, estimating forest parameters in large areas using a traditional inventory approach combined with satellite data analysis can improve the spatial estimates of forest inventory data, and hence be useful for sustainable forest management and natural resources assessment. However, extracting practical information from satellite imagery for such purpose is a challenging task mainly because of insufficient knowledge linking forest inventory data to satellite spectral response. Here, we present the use of a cost-free Landsat-7 Enhanced Thematic Mapper Plus (ETM+) in order to explore whether it is possible to combine all available optical bands from a specific sensor for improving forest parameter spatial estimates, based on fieldwork at Lahav and Kramim Forests, in the Israeli Northern Negev. A generic strategy, based on morphological structuring element, convex hall and spectral band linear combination algorithms, was developed in order to extract the mathematical dependencies between the forest inventory measurements and linear combination sets of Landsat-7 ETM+ spectral bands, which yields the highest possible correlation with the forest inventory measured data. Using the mathematical dependency functions, we then convert the entire Landsat-7 ETM+ scenes into forest inventory parameter values with sufficient accuracy and tolerance errors needed for sustainable forest management. The root mean square error obtained between the measured and the estimated values for Lahav Forest are 0.70 cm, 0.29 m, and 1.48 m³ ha⁻¹ for the mean DBH, H, and V, respectively, and for Kramim forest are 0.61 cm, 0.70 m, and 6.31 m³ ha⁻¹, respectively. Furthermore, the suggested strategy could also be applied with other satellites data sources.
... L and C bands) are still sensitive to forest parameters; however, Lband exhibits some saturation for high WV values (> 400 -500 m 3 /ha) and C band, with its scarce penetration inside the vegetation cover, is only sensitive to the first layers of the crowns. The latter frequency is more suitable for classification purposes and investigation of forest features related to degradation [Dobson et al., 1992;Le Toan et al., 1992;Beaudoin et al., 1994;Israelsson et al., 1994;Rauste et al., 1994;Paloscia et al., 1999]. The use of multi-frequency and multi-polarization approaches can therefore provide accurate results, being able to characterize the different layers of the trees. ...
This work aims at investigating the potential of L (ALOS/PALSAR) and C (ENVISAT/ ASAR) band SAR images in forest biomass monitoring and setting up a retrieval algorithm, based on Artificial Neural Networks (ANN), for estimating the Woody Volume (WV, in m 3 /ha) from combined satellite acquisitions. The investigation was carried out on two test areas in central Italy, where ground WV measurements were available. An innovative retrieval algorithm based on ANN was developed for estimating WV from L and C bands SAR data. The novelty consists of an accurate training of the ANN with several thousands of data, which allowed the implementation of a very robust algorithm. The RMSE values found on San Rossore area were ≅40 m 3 /ha (L band data only), and 25-30 m 3 /ha (L with C band). On Molise, by using combined data at L and C bands, RMSE<30m 3 / ha was obtained.
... EO techniques have already been applied to map forest cover extent, as well as changes in forest cover caused by clear-cut logging and forest fires (Fraser & Li, 2002;Fraser, Li, & Cihlar, 2000;Healey, Cohen, Yang, & Krankina, 2005;Kasischke, Bourgeauchavez, French, Harrell, & Christensen, 1992;Lozano, Suarez-Seoane, Kelly, & Luis, 2008;Rignot, Salas, & Skole, 1997;Thiel, Thiel, & Schmullius, 2009;White, Wulder, Brooks, Reich, & Wheate, 2005;Yatabe & Leckie, 1995). Current EO techniques applied for the estimation of forest stem volume still struggle with problems related to saturation and considerable uncertainties (Dobson et al., 1992;Israelsson, Askne, & Sylvander, 1994;Santoro, Eriksson, Askne, & Schmullius, 2006). range spectral filter (Santoro, Werner, Wegmüller, & Cartus, 2007;Wegmüller, 1998). ...
The full potential of ALOS PALSAR L-band interferometric (InSAR) coherence data for the estimation of forest growing stock volume (GSV) in the boreal forest has rarely been investigated. Moreover, ALOS PALSAR backscatter and InSAR coherence have yet to be used together to delineate GSV. Due to the observation strategy and the high acquisition success rate over Eurasia, a large amount of high quality ALOS PALSAR L-band data is available over Siberia. Consequently, this paper investigates the capability of ALOS PALSAR backscatter and InSAR coherence for the estimation of GSV in Central Siberia, Russia. The potential of backscatter and coherence are directly compared using the same inventory data. Altogether, 87 PALSAR images are used and eleven forest inventory sites are investigated.
Based on this large dataset it was observed that InSAR coherence acquired in frozen conditions offers the highest potential for GSV estimation. The saturation level for single coherence images was on average 230 m³/ha, with an average R² between coherence and GSV of 0.58. PALSAR backscatter acquired in unfrozen conditions could also estimate GSV; however, the saturation levels (75–100 m³/ha) and the average R² (0.42–0.48) were lower. HV backscatter offered only a slightly greater potential than HH backscatter.
A simple inversion approach aiming at the delineation of forest GSV maps based on the multitemporal SAR data was developed and applied to five forest inventory sites. This approach combines HV backscatter data acquired in unfrozen conditions and InSAR coherence data acquired in frozen conditions. In general, the produced maps feature a corrected relative RMSEcorr of < 30% which was similar to the accuracy of the forest inventory data. The R² between inventory data and SAR data based maps varied between 0.54 and 0.83.
... The radar backscattering coefficient is correlated with forest biomass and stem volume (Le Toan et al. 1992;Israelsson et al., 1994;Kasischke et al., 1994, Dobson et al., 1995. The sensitivity of Synthetic Aperture Radar (SAR) data to forest stem volume increases significantly as the radar wavelength increases (Israelsson et al., 1997). ...
... SAR data is widely used for forest cover mapping [1], forest disturbance mapping (e.g. logging, forest fire, and wind damage) [2] [3] [4] and forest biomass assessment [5] [6] [7]. Lower frequencies proved to be of particular adequacy. ...
This paper provides an initial assessment of the qualification of polarimetric parameters for forest cover mapping in Siberian boreal forest. This investigation is carried out in the framework of JAXA's Kyoto and Carbon Initiative. The impact of the additional polarimetric information will be related to the backscattering intensities derived with the respective polarimetric acquisitions. The dataset consists of two Level 1.1 PLR scenes acquired during the calibra-tion/validation phase and is thus not determined by the schedule of the ALOS acquisition strategy. First results show the capability of polarimetric parameters to extend the intensity based data set in terms of a higher overall separability of forest and non forest and to some extend the feasibility of subdividing both basic classes.
... SAR coherence was demonstrated being a valuable estimator of forest biomass (Israelsson at al., 1994, Leckie & Ranson, 1998). Consequently, this technique can be also employed for forest cover and forest cover change mapping ( Yatabe et al., 1995, Santoro et al., 2006). ...
This paper explains the role and the main objectives of the GSE Forest Monitoring (GSE FM) Service Portfolio Evolution. This task supports the cardinal requirements of sustainability of the GMES (Global Monitoring for Environment and Security) services. In order to satisfy the needs of the stakeholder the requests for better and more complete information on environment and security the GMES services must continuously improved. The Service Portfolio Evolution is one of the key inputs for improved service provision. Throughout the entire services of the GSE FM Service Portfolio the interactive involvement within the process of production (in-situ measurements, data pre-processing, data classification, product accuracy assessment etc.) is the most cost and time efficient factor and is therefore be treated with high priority to research activities. The continuous identification of new R&D therefore supports the sustainability of the GMES services and establishes common standards for all Service Providers and users.
... Für die Erfassung von aktuellen Waldbeständen und Waldflächenveränderungen (Abholzung, Waldbrände, Sturmschäden, Insektenbefall etc.) sowie Biomasseableitungen hat sich die Eignung von Radardaten innerhalb der letzten Jahre erwiesen (LECKIE & RANSON 1998, YATABE & LECKIE 1992, RIGNOT et al. 1997, DOBSON et al. 1992, ISRAELSSON et al. 1994, SANTORO et al. 2006. Insbesondere für die Aktualisierung von Forstinventurdatenbanken stellte sich heraus, dass Radardaten mit geringer Frequenz von besonderer Bedeutung sind. ...
Um großflächige Walduntersuchungen durchzuführen, hat sich in den letzten Jahren der Einsatz von Fernerkundungsdaten, speziell von Radardaten stark bewährt, um aktuelle Waldbestände und Waldflächenveränderungen (Abholzung, Waldbrände, Sturmschäden, Insektenbefall etc.) zu erfassen und zu überwachen. In dieser Arbeit konnte gezeigt werden, dass eine deutliche Unterscheidung zwischen Wald und Nichtwaldgebieten gemacht werden kann, wenn ALOS PALSAR Sommerintensitäten und ALOS PALSAR Winterkohärenzen zum Einsatz kommen.
... [16,17,20,28,29,30,31]. It has been showed that the shorter wavelengths such as Xband and C-band are primarily scattered in forests by small tree components, typically the foliage, twigs and branches of the upper canopy [21,32,33,34]. Short wavelengths penetration being limited to the surface of the forest canopies, some authors showed furthermore their potential of estimating other forest biophysical parameters such as the forest canopy cover or forest species composition [35]. ...
The application of remote sensing techniques in the retrieval of forest stand proprieties, especially biomass and stem volume, has been intensively investigated during the last few years. New methods using SAR and Optical systems have been pointed out. However, these instruments present intrinsic limitations which prevent operational biomass retrieval while using them separately. In order to overcome SAR and Optical system restrictions, with respect to biomass mapping, the present study will assess the potential of fusion of these two different sources of information. The Thuringian Forest located in the center of Germany has been chosen for the investigations. An extensive set of L-Band data (ALOS PALSAR) and X-Band data (TerraSAR-X) has been acquired in various sensor configurations (acquisition modes, polarizations, incident angles). High resolution optical data (RapidEye and Kompsat2) has been also acquired and a forest inventory with a great deal of forest stands has been delivered. The study will first concentrate on the pre-processing of the data. A data analysis will then be carried out and preliminary results will be discussed with a view to the fusion of SAR and Optical information. The fusion of SAR and Optical systems has the potential to give new possibilities in terms of temporal and spatial transferability. In this context, it will be expected to increase the biomass map accuracy, its spatial extension and its updated frequency.
... basal area) and cannot be measured straightforward on the field, but require allometric calculation. Much remote sensing research has focused on the extraction of forest stand parameters using optical (Franklin, 1986;Stenback and Congalton, 1990;Ardo, 1992;Curran et al., 1992;Cohen et al., 1995;Trotter et al., 1997;Hyyppä et al., 2000;Franco-Lopez et al., 2001;Lu et al., 2004;Muukkonen and Heiskanen, 2005) and microwave radar data (Le Toan et al., 1992;Israelsson et al., 1994;Rauste and Hame, 1994;Fransson and Israelsson, 1999;Santoro et al., 2001;Sun et al., 2002). ...
... At L-band, the effect of climate conditions at acquisition has been reported in 4-6, 8, 12 . In 6,7,13,14 it has been shown that the backscatter in boreal forests saturates between 150 and 225 m 3 /ha, thus being more suitable than X-and C-band for stem volume retrieval. ...
In this paper a methodology for growing stock volume retrieval in boreal forests is devised. The methodology is developed at stand level and uses a multi-temporal set of spaceborne SAR data acquired at L-band by the Japanese satellite JERS-1 over an interval of four years. The test area consists of four large forest compartments, being part of the forest enterprise of Bolshe-Murtinsky, Siberia. Although two compartments were characterized by marked topography, the backscatter did not seem to be affected, in particular in dense forests. Seasonal effects were more relevant. SAR imagery acquired under dry-unfrozen conditions consis-tently showed the largest dynamic range and the absence of saturation in the interval 0-350 m 3 /ha. By means of a semi-empirical model the inversion of backscatter measurements was performed and growing stock vol-ume was estimated. Dry-unfrozen conditions were the most suitable for the retrieval, although the effect of spatial heterogeneities within the compartments increased the error. For changing weather conditions, satura-tion at 200 m 3 /ha limited the retrieval to sparse forests. Frozen conditions were the least suitable. Multi-temporal combination of single-image estimates helped in increasing the accuracy. Using a rough correction for the errors in the ground-truth, the growing stock volume estimates were reasonably accurate and in line with studies carried out at other boreal sites, being mainly affected by spatial differences of the backscatter in the test area and the properties of the growing stock volumes used for model training. This study shows the importance of L-band backscatter for the retrieval of growing stock volume in the boreal belt, thus being an important aspect for the forthcoming ALOS mission.
... Most researchers attempting to extract information content from radar data using digital image processing have confined their study to a limited range of incidence angles, typically spanning approximately 10 degrees (Rignot et al., 1994a;Israelsson et al., 1994;Dobson et a]., 1992;Hussin et al., 1991). It is argued that the effect of incidence angle in this limited range does not cause significant changes in backscatter characteristics. ...
Three image-based methods for correcting the effect of changes in incidence angle on backscatter data are proposed and evaluated for AirSAR data of a coastal tropical savanna landscape in Australia's Northern Territory. These correction methods require little field knowledge and do not assume a linear relationship between sAR backscatter and incidence angle. The correction methods are applied to five independent components of the Stokes Matrix. The results are evaluated using an existing land-cover classification to extract mean backscatter values for individual land-cover classes before and after the correction procedures. It is shown, for the vegetation communities involved, that the slope method provides a successful correction for the amplitude components of the backscatter data. The real and imaginary parts of the co-polarized returns are best corrected using the LuT method. The proposed correction procedure allowed the use of a maximum-likelihood classification for the discrimination of ten land-cover types with an overall accuracy of 87 percent across the complete swath width of the AirSAR data.
... This means that the models derived in these papers usually have little or no application outside the studied test site. This is an obvious disadvantage when global biomass mapping is [5], [7]. NOTE, THAT IN THE CASES OF [10], [12], AND [13] FORWARD MODELS ARE DEFINED. ...
A new biomass retrieval model for boreal forest using polarimetric P-band synthetic aperture radar (SAR) backscatter is presented. The model is based on two main SAR quantities: the HV backscatter and the HH/VV backscatter ratio. It also includes a topographic correction based on the ground slope. The model is developed from analysis of stand-wise data from two airborne P-band SAR campaigns: BioSAR 2007 (test site: Remningstorp, southern Sweden, biomass range: 10–287 tons/ha, slope range: 0–4
) and BioSAR 2008 (test site: Krycklan, northern Sweden, biomass range: 8–257 tons/ha, slope range: 0–19
). The new model is compared to five other models in a set of tests to evaluate its performance in different conditions. All models are first tested on data sets from Remningstorp with different moisture conditions, acquired during three periods in the spring of 2007. Thereafter, the models are tested in topographic terrain using SAR data acquired for different flight headings in Krycklan. The models are also evaluated across sites, i.e., training on one site followed by validation on the other site. Using the new model with parameters estimated on Krycklan data, biomass in Remningstorp is retrieved with RMSE of 40–59 tons/ha, or 22–33% of the mean biomass, which is lower compared to the other models. In the inverse scenario, the examined site is not well represented in the training data set, and the results are therefore not conclusive.
... Radar remote sensing of forested areas have been investigated since many years. At high frequencies (C, L and P band) these investigations show some limitations as for example in estimating forest biomass parameters, and also a saturation of the backscattered signal [1][2][3] and [4]. At VHF frequencies, as with the Swedish CARABAS SAR system which operates in the 20-90 MHz band [5,6] and [7], the electromagnetic wave crosses the canopy and then we have to consider the interaction of the wave with the trunks and the ground. ...
Abstract—A forest made of an infinite biperiodic array of trees over a lossy ground, is illuminated by a linearly polarized electromagnetic plane wave in the range of 20 to 90 MHz. Due to the ratio of the wavelength to the array period, only the specular mode is propagative. Therefore, a reflection coefficient is computed and not a backscattering coefficient. It is obtained by means of a full wave approach, based on an integral representation of the electric field. This approach takes into account all possible interactions between each component,of the medium,as well as ground penetration and provides full information on the phase of the scattered field. Two models of the forest are developed, the two layers one where trees are separated and the four layers one whee the canopy is replaced by an equivalent homogeneous,medium. The low frequency (VHF) used here make this homogenization possible and allows one to consider trees with simple shape, the wave being unable to sense details of a tree.
... It has been shown that in many cases there is a correlation between backscatter intensity and forest biomass (e.g. Dobson et al. 1992, Israelsson et al. 1994, Rignot et al. 1994). An important factor aVecting the backscatter is the structure of the trees. ...
The use of spaceborne synthetic aperture radar (SAR) systems to estimate stem volume and biomass in boreal forests has shown some promising results, but with saturation of the radar backscatter at relatively low stem volumes and limited accuracy of stem volume estimation. These limitations have motivated evaluation of more advanced methods, such as interferometry. The results presented in this study show that ERS interferometry, under favourable conditions, may be used to estimate stem volume at stand level with saturation level and accuracy useful for operational forestry management planning in boreal forests. Five interferograms were analysed, covering a test site located in the central part of Sweden with stem volume in the range of 0-305 m3 ha-1. The best interferogram showed a linear relationship between stem volume and coherence with a root mean square error (RMSE) of approximately 26 m3 ha-1, corresponding to 20% of the average stem volume, throughout the range of stem volume. No saturation was observed up to the maximum stem volume. However, the sensitivity of coherence to stem volume varied considerably between the interferograms. Finally, four SPOT XS images were evaluated and compared with the stem volume estimations obtained from the interferograms, resulting in a relative RMSE of about 24% of the stem volume, for the best case. The estimation of stem volume using coherence data was found to be better than optical data for stem volumes exceeding about 110 m3 ha-1. The statistical analysis was performed using linear regression models with cross-validation.
... Due to the scarcity of soil information, lithological and geomorphologic information is expected to serve as a substitute. In the absence of potentially appropriate radar systems (Israelsson et al. 1994), the best possible use must be made of optical systems to classify vegetation, as demonstrated by Foody and Hill (1996j. With respect to biogeochemical process studies (see below), information about the canopy chemistry of thus classified forests is desirable, with an emphasis on, but not limited to, nitrogen and lignin. ...
Enormous meandering rivers are the most remarkable fluvial feature of the Amazon landscape, but these rivers are only the largest component of a much denser network of streams which finely dissects and drains the basin. In terms of combined length and total amount of lotic habitat, streams dominate over their more visible downstream counterparts; this dominance is especially dramatic for first- and second- order streams which alone may account for greater than 80% of total channel length in meso-scale Amazon drainag basins (Table 12.1). The flow of Amazon streams emerges directly from the extensive forests and savannas that compose the basin. Bio- geochemical cycles in streams are thus intricately associated with processes operat- ing in adjoining riparian and upland ecosys- tems. Terrestrial processes regulate the input of organic and inorganic species to stream systems, and the chemistry of inflowing
... Recently, much remote-sensing research has focused on the extraction of forest stand parameters through correlation or regression analysis to examine relationships between spectral response and structural factors of coniferous forest, including basal area, biomass, crown closure, diameter at breast height (DBH), tree height, vegetation density, and leaf area index (LAI) using optical sensor data such as Landsat Thematic Mapper (TM) images (Franklin, 1986; Horler and Ahern, 1986; Peterson et al., 1986 Peterson et al., , 1987 Spanner et al., 1990; Stenback and Congalton, 1990; Lathrop and Pierce, 1991; Ardo, 1992; Curran et al., 1992; Cohen et al., 1995; Gemmell, 1995; Kimes et al., 1996; Trotter et al., 1997; Turner et al., 1999; Eklundh et al., 2001; Franco-Lopez et al., 2001). Similar forest parameter data have been acquired through analysis of microwave radar images such as synthetic aperture radar (SAR) (Israelsson et al., 1994; Rauste and Hame, 1994; Harrell et al., 1995 Harrell et al., , 1997 Fransson and Israelsson, 1999; Kurvonen et al., 1999; Santoro et al., 2001; Castel et al., 2002; Sun et al., 2002). Due to the important roles of moist tropical forests in global warming, biodiversity, and ecosystems , research using remotely sensed data to measure selected properties of tropical forest stand parameters has increasingly attracted interest during the past decade (Cook et al., 1989; Sader et al., 1989; Wu, 1990; Lucas et al., 1993; Foody and Curran, 1994; Nelson et al., 2000; Steininger, 2000; Lu, 2001; Tetuko et al., 2001; De Wasseige and Defourny, 2002; Drake et al., 2002; Lu et al., 2002a; Santos et al., 2002). ...
Estimation of forest stand parameters such as aboveground biomass in a large area using remotely sensed data has considerable significance for sustainable management and utility of natural resources. In practice, selecting suitable image data for such purposes remains difficult due to a poor understanding of forest stand parameters and remote-sensing spectral response relationships, particularly in moist tropical regions. This paper explores relationships between forest stand parameters and Landsat Thematic Mapper (TM) spectral responses through analyses of three study areas in the eastern Amazon basin (Altamira, Bragantina, and Ponta de Pedras). Six TM bands and many vegetation indices are examined through integration of spectral responses and field vegetation inventory data. Pearson’s correlation coefficients are used to interpret relationships between forest stand parameters and TM data. This study concludes that single band TM5 and linear transformed indices such as PC1 (the first component in a principal component analysis), KT1 (brightness of the tasseled cap transform), and albedo are most strongly correlated with forest stand parameters, somewhat independent of biophysical environments. Many vegetation indices that use TM4 and TM3 data, such as the atmospherically resistant vegetation index, the atmospheric and soil vegetation index, and the normalized difference vegetation index, are weakly correlated with selected forest stand parameters. In contrast, vegetation indices using band TM5 data improve correlations with selected forest stand parameters in Altamira forests that are characterized by a complex stand structure. Forest stand structure and associated canopy shadow affect the forest stand parameters and TM spectral response relationships. This paper provides a better understanding of relationships that have a potential of being important for developing stand parameter estimation models and for improvement of vegetation classification accuracy.
The objective of this study is to investigate a unique data set consisting of SAR images acquired at P–band, X–band and VHF over a forested area situated in the southern part of Sweden. The Backscattering Coefficients (s o) for an intensity analysis and the Mean Contrast from a Grey Level Cooccurence Matrix for a texture analysis of the imagery have been related to ground truth stem volume measurements by mean of linear regression. Moreover, a new textural method based on the distance dependence of the standard deviation of the radar backscattering coefficients was used.The results obtained show the advantage of using lower frequencies for stem volume retrieval with both intensity and texture analysis. The higher correlation coefficients for the intensity analysis and the lesser demand for both storage and computing time compared to texture analysis suggests using the former as investigation method for forest biomass retrieval.
Radar backscatter variations that occur because of incidence angle effects constrain the application of Scanning Synthetic Aperture Radar (ScanSAR) data for sea ice monitoring and observations. In this paper, a class-based correction is proposed for normalizing each class in ScanSAR data to a nominal incidence angle. Two tested sea ice synthetic aperture radar (SAR) data sets were acquired: a data set for the Gulf of Saint Lawrence, which was obtained by the RADARSAT-2 satellite, and a data set for the Bohai Sea, which was obtained by the ENVISAT Advanced Synthetic Aperture Radar. An unsupervised classification is performed on each image block prior to normalization, and the incidence angle range of each image block is approximately 5°. Because the distribution of the backscatter coefficients in the azimuth band is discrete and nonlinear, the class-based locally linear mapping (LLM) technique is implemented, based on the assumption that a small quantity of sorted backscatter coefficients is locally linear. This algorithm is a transplantable and easily applied method that requires limited ground data, and it is also a semiautomated technique because nearly all of its parameters can be adaptively determined during the image analysis. The results demonstrate that LLM-corrected ScanSAR images appear to have more detailed textures, and the natural signal variability in the radar data is preserved, which indicates that the LLM produces better results compared with the histogram-based-alike (HIST-alike) technique when correcting the incidence angle in the sea ice SAR data. The results of the data analysis in this paper show that the width of the azimuth band should be selected based on the extent of variation in the incidence angle, and the reference band can be calculated based on the maximum interclass distance principle. The intercomparisons also reveal that the proposed algorithm can improve the accuracy of supervised classifications.
In this paper summer intensity and winter coherence images are used for large scale forest monitoring. The intensities (FBD HH/HV) have been acquired during summer 2007 and feature the K&C intensity stripes [1]. The processing consisted of radiometric calibration, orthorectification, and topographic normalisation. The coherence has been estimated from interferometric pairs with 46-days repeat-pass intervals. The pairs have been acquired during the winters 2006/2007 and 2007/2008. During both winters suited weather conditions have been reported. Interferometric processing consisted of SLC co-registration at subpixel level, common-band filtering in range and azimuth and generation of a differential interferogram, which was used in the coherence estimation procedure based on adaptive estimation. All images were geocoded using SRTM data. The pixel size of the final SAR products is 50 m x 50 m. It could already be demonstrated, that by using PALSAR intensities and winter coherence forest and nonforest can be clearly separated [2]. By combining both data types hardly any overlap of the class signatures was detected, even though the analysis was conducted on pixel level and no speckle filter has been applied. Thus, the delineation of a forest cover mask could be executed operationally. The major hitch is the definition of a biomass threshold for regrowing forest to be distinguished as forest.
Synthetic aperture radar (SAR) is a technology to generate high-resolution radar images for ground surveillance and mapping. By exploiting the lower VHF band, the attenuation reduces dramatically in forest areas compared to microwaves and new applications become feasible. Examples of these are detection of targets obscured by vegetation, stem volume estimation of forest stands, and generation of "bald earth" digital elevation models in forested areas. This thesis presents research results of importance for ultra-wideband VHF-band SAR, including interferometric technique. The processing steps and the results for stem volume retrieval are described in an early stage study of VHF-band SAR data. Low attenuation and backscattering levels were observed and thus a potential for forest inventory applications was recognized. Novel results of generating interferometric fringe images using VHF-band SAR data in a forest area are treated, and parts of the processing steps are discussed. The effects of the motion compensation for the interferometric coherence are investigated. The conclusion is that the coordinate system of individual SAR images should be selected as close as possible to the actual track heading. Further investigation of motion compensation effects in SAR processing is presented for quantitative geometrical error estimates of targets elevated from a nominal focusing height. An extensive investigation of bald earth height estimation in a boreal coniferous forest area is carried out and the choice of baseline is discussed. A wide range of baselines are applicable and give a similar height estimate quality. For a large area of 1 × 2 km2 an rms deviation of about 2 meters from ground truth data was obtained. Especially good results are obtained for small areas consisting of dense forest with strong backscattering and continuous high coherence, i.e. with ground truth rms deviation of less than 1 meter. As a final issue, angular interpolation is evaluated in a time domain SAR algorithm, based on fast factorised backprojection for generating high quality images. The results show that an image quality in parity with a global backprojection (GBP) reference image can be reached, using an adequate sampling rate and interpolation kernel size, while still gaining the benefit of a large reduction in the number of operations required. Using a cubic interpolator and a sampling rate of two times the maximum angular bandwidth, 2Bθ, and applying a Hamming window, the ISLR and PSLR along the one-dimensional azimuth direction is measured to -58 dB and -42 dB, respectively. Increasing the angular sampling rate to 8Bθ, results in an ISLR and PSLR of-60 dB and -48 dB, respectively, which equals the result obtained by the GBP reference algorithm.
Based on a quantitative study of the incident angle effect of a wide-swath synthetic aperture radar (SAR) image, this paper proposes a class-based correction method for such an effect. The method achieves sampling by using watershed segmentation and regional labeling technology as well as a class-based radiation correction of land-cover backscatter values based on the cosine Lambert's law estimated through linear regression. Experimental results of the Envisat Advanced SAR (C-band horizontal-transmitting, horizontal-receiving polarization) data show that an incident angle effect is exerted on radar backscatter; the higher the moisture content of the land-cover type is, the more obvious the incident angle effect is. The correction of incident angle effect with this proposed algorithm is better than order cosine correction.
Methods to estimate forest biomass change have been investigated using experimental P-band synthetic aperture radar (SAR) data from the recent airborne campaigns BioSAR 2007 and BioSAR 2010 conducted over a hemiboreal test site in southern Sweden. Regression models based on backscatter change were developed using reference biomass change maps derived from high-density laser scanning data. Different regression models were developed for linear, square root, and logarithmic biomass change scales. The models were compared to the change maps based on laser data using twofold cross-validation, and estimation errors were evaluated using six 80 m by 80 m plots with detailed in situ measurements. The results indicate that the root-mean-square error of biomass change estimates based on P-band SAR backscatter data is about 15% or 20 t/ha. This suggests that not only clear-cuts but also growth and thinning can be measured. Simulations were performed in order to evaluate the possibility of using a spaceborne P-band SAR for measurements of forest biomass change. The simulations show that, with 64 equivalent number of looks (ENL) and a 50% change in biomass, it is possible to correctly indicate whether the forest has gained or lost biomass. Similarly, for a biomass loss of more than 75%, a correct indication of the sign of biomass change can be achieved with only 8 ENL.
Numerous studies demonstrated the potential of interferometric coherence for forest stem volume estimation in boreal forests. Coherence derived from winter images acquired at frozen conditions proved being favourable against unfrozen conditions. This also applies to PALSAR coherence, although featuring relatively large temporal baseline of at least 46 days. However, when using data acquired at unfrozen conditions, a large spread of coherence was observed at all stem volume levels. This scatter negatively effects the correlation of stem volume and coherence, and thus prevents using summer coherence for stem volume estimation. Besides environmental conditions, the different tree geometries can be assumed having an impact on the magnitude of decorrelation. So far, this impact has rarely been investigated in boreal forests. This paper presents the results of a study investigating the impact of tree species on PALSAR coherence. The results show increased scatter of coherence in summer. They also show that the coherence of dense forest is larger in summer than in winter. Eventually, a slight impact of the tree species on coherence was observed. This impact is larger in summer.
Multi-temporal ERS-1 C- and JERS-1 L-band Synthetic Aperture Radar (SAR) data were analysed to determine the relationship between radar backscattering and forest stem volume. The test site is located within the boreal conifer belt in northern Sweden with stem volume in the range of 0-300m3 ha-1. The statistical analysis was carried out using the water cloud model and various linear regression models. It was confirmed that the L-band of JERS-1 shows a consistently higher sensitivity to stem volumes in boreal forests with a larger backscatter contrast between nonforested areas and areas having dense tree cover in comparison with the C-band of ERS-1. The saturation levels for the ERS-1 and JERS-1 SAR sensors were estimated at 64m3 ha-1 and 143m3 ha-1, respectively. Furthermore, the temporal analysis shows that the backscattering coefficient can differ by approximately three decibels, atmost, between acquisitions for a given stem volume. However, a major finding is that the temporal variation could be modelled as an additive effect that is constant throughout the range of stem volume. For operational forestry purposes this would imply that stem volume can be predicted after calibration of the radar response using clear felled areas as reference targets. The multi-temporal JERS-1 dataset was used to develop a radarbased model of stem volume. The model was tested on a validation dataset resulting in a correlation coefficient of 0.78 between SAR estimated stem volume and ground data.
Forests have been proposed as a means of reducing atmospheric carbon dioxide levels due to their ability to store carbon as biomass. To quantify the amount of atmospheric carbon sequestered by forests, biomass and density estimates are oven needed. This study develops, implements, and tests an individual tree-based algorithm for obtaining forest density and biomass using orthophotographs and small footprint LiDAR imagery. It was designed to work with a range of forests and image types without modification, which is accomplished by using generic properties of trees found in many types of images. Multiple parameters are employed to determine how these generic properties are used. To set these parameters, training data is used in conjunction with an optimization algorithm (a modified Nelder-Mead simplex algorithm or a genetic algorithm). The training data consist of small images in which density and biomass are known. A first test of this technique was performed using 25 circular plots (radius = 15 m) placed in young pine plantations in central Virginia, together with false color orthophotograph (spatial resolution = 0.5 m) or small footprint LiDAR (interpolated to 0.5 m) imagery. The highest density prediction accuracies (r2 up to 0.88, RMSE as low as 83 trees/ha) were found for runs where photointerpreted densities were used for training and testing. For tests run using density measurements made on the ground, accuracies were consistency higher for orthophotograph-based results than for LiDAR-based results, and were higher for trees with DBH >=10cm than for trees with DBH >=7 cm. Biomass estimates obtained by the algorithm using LiDAR imagery had a lower RMSE (as low as 15.6 t/ha) than most comparable studies. The correlations between the actual and predicted values (r2 up to 0.64) were lower than comparable studies, but were generally highly significant (p
In this study, we investigated the potential of polarimetric synthetic aperture radar (PolSAR) data for the estimation of stem volume in tropical forests. We used calibrated L-band, high incidence angle data from the airborne system SAR-R99B, acquired over an experimental area in the Tapajós National Forest, Pará, Brazil. To evaluate the potential of PolSAR data for this application we used regression analysis, in which first-order models were fit to predict stem volume per hectare, as determined from field measurements. Unlike previous studies in tropical forests, the set of potential explanatory variables included a series of PolSAR attributes based on phase information, in addition to power measurements. Model selection techniques based on coefficient of determination (R ) and mean square error (MSE) identified several useful subsets of explanatory variables for stem volume estimation, including backscattering coefficient in HH polarization, cross-polarized ratio, HH-VV phase difference, polarimetric coherence, and the volume scatter component of the Freeman decomposition. Evaluation of the selected models indicated that PolSAR data can be used to quantify stem volume in the study site with a root mean square error (RMSE) of about 20–29 m ha, corresponding to 8–12% of the mean stem volume. External validation using independent data showed average prediction errors of less than 14%. Saturation effects in measured versus modelled volume were not observed up to volumes of 308 m ha (biomasses of ∼357 Mg ha). However, no formal assessment of saturation was possible due to limitations of the volume range of the dataset.
The spectral separability of thirteen topical vegetation classes, including twelve forest types, was assessed. Although the thirteen classes could not be classified to a high accuracy the results of a set of supervised and unsupervised classifications revealed that three groups of classes were highly separable; a classification of the three groups by a discriminant analysis had an accuracy of 92·20 per cent. These three spectrally separable groups also corresponded closely to ecological groups identified from an ordination of data on tree species contained within a detailed ground data set. On the basis of the class separability analyses the three spectrally separable groups were mapped, with an accuracy of 94·84 per cent, from Landsat TM data by a maximum likelihood classification. It was apparent that some of the errors in this classification could be resolved through the use of contextual information and ancillary information, particularly on topography.
New forest, inventory methods must be developed in order to create good conditions For decision-making regarding ecological and economical issues in forestry. There are good field measurement methods to use, but they are often very expensive. Coherent ail radio band sensing (CARABAS) is a newly developed synthetic aperture radar (SAR) sensor. It differs significantly from the earlier SARs by using longer wavelengths. The CARABAS sensor is more adapted to the scatterers of interest in the forest. due to its longer wavelengths. In this study, CARABAS imagery is compared with forest tree volume. Regression analysis was used to relate radar backscattering to forest tree volume. Due to the large range of incidence angle (45 degrees-65 degrees), the CARABAS image had to be radiometrically relative-calibrated. Radar backscattering from five forest stands with similar volume contents were plotted against the distance from the sensor, The plot revealed a linear relationship between these variables. By linear regression on that material the other pixels were relative-calibrated in the image. Finally, radar backscattering was related to forest stand volumes by using linear regression. The results showed that the backscattering component of the CARABAS imagery is highly correlated to forest tree volume (R-2 = 74.9%). In this material, there seemed to be no saturation level of the backscattering component up to 300 m(3) ha(-1).
The paper analyses seasonal effects on L-band backscatter in boreal forests and the implications for stem volume retrieval (JERS-1 mission). As test sites, the estate of Kättböle, Sweden, and two compartments in Bolshe-Murtinsky, Siberia, were considered. The in-situ measured stem volumes ranged from 5 to 350 m3/ha in Kättböle and to 400 m3/ha in Bolshe-Murtinsky, at stand level. For each site nine SAR images were available. Forest backscatter strongly depended on seasonal conditions. With respect to other seasons, in frozen conditions the dynamic range was smaller and the forest backscatter at least 3 dB lower. When precipitation occurred, the backscatter showed saturation. In Kättböle, no saturation was found in images acquired at dry/unfrozen conditions. By means of a semi-empirical model, a regression between stem volume and backscatter was performed. Stem volume was then retrieved for an independent set of backscatter measurements. Images acquired at dry/unfrozen conditions showed a relative RMS error of around 30 % for the images acquired over Kättböle. At both sites the retrieval error was higher for other weather conditions, around 50%. When dry/unfrozen conditions occurred, multi-temporal combination of stem volume estimates showed the smallest error (22%). Hence, for boreal forest monitoring L-band images acquired at dry/unfrozen conditions should be used.
Radar imaging is a valuable tool for the monitoring and management of tropical ecosystems. One of the obstacles to its successful usage is the variation of the radar backscatter with incidence angle caused by the side-looking mode of the sensor. Available research has modelled the effect of changes in incidence angle for soil, grass and scattering mechanisms in tree-covered areas. The effect varies depending on the scattering mechanism. To model the radar backscatter behaviour of a natural landscape, therefore, implies precise knowledge of the land-cover composition. A method for estimating and correcting the effect of changes in look angle on backscatter data is consequently proposed; it requires little field knowledge and encompasses the effects on most land-cover types in the study area. The method is based on the assumption that each line in azimuth direction contains a similar composition in regard to land-cover types. The backscatter frequency distribution of each azimuth line can then be employed to model and correct for the effect of variation in incidence angle. The results are evaluated using an existing land-cover classification to extract mean backscatter values for individual land-cover classes before and after the correction procedure. The correction procedure is shown to successfully adjust backscatter intensities to a nominal incidence angle for the vegetation communities in a coastal tropical savanna landscape in Australia's Northern Territory.
Due to the important role of forests in the global carbon cycle, it’s necessary to pay more
attention to these regions. Among forests biophysical parameters, Biomass estimation as
the most influencing factor on the ecosystem is the most important. Due to limitations of
traditional forest inventory technique to measure biomass, remote sensing has been
widely used for biomass estimation in the past decade. In areas with complex structure,
various species and multilayer vegetation cover, using microwave images would be the
only solution because of the limitation of optical images. According to recent studies
about biomass estimation with microwave images, the interferometry technique yields the
best results but still is limited and not tested in all forest types. Because using microwave
images has its own problems, focusing on the development of the new approaches which
can handle multi source data, identify and decrease uncertainty sources in biomass
estimation is required in future studies.
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.
An integrated remote sensing/field ecology project linked the use of synthetic aperture radar (SAR) and aerial photography to studies of landscape spatial heterogeneity and bird community ecology. P-, L-, and C-band SAR data, collected over a section of Kakadu National Park in Australia's Northern Territory during the Joint NASA/Australia DC-8 data acquisition campaign, were analyzed in light of field data integrating vegetation structure and floristics with bird abundances across a heterogeneous study site. Results indicate that SAR data are able to discern structural differences relevant to bird habitat quality within floristically homogeneous stands, while multispectral sensors successfully identified floristic differences among habitat types. Simplifying indices of bird diversity showed ambiguous changes across the site; however, the abundances of individual species were observed to change significantly across both floristic and structural gradients. These results suggest that efforts to map bird diversity should focus on species-specific habitat relationships and that some measure of vegetation structure is needed to understand bird habitat. The approach employed here advances the use of SAR data in the three-dimensional mapping of animal habitats from remotely sensed data, and extends current capabilities for mapping and modeling large-scale patterns in the distribution of biological diversity.
At present, the greatest source of uncertainty in the global carbon cycle is in the terrestrial ecosystems. In order to reduce these uncertainties it is necessary to provide consistent and accurate global estimates of the world forest biomass. One of the most promising methods for obtaining such estimates is through polarimetric SAR backscatter measurements at low frequencies. In this paper, the relation between polarimetric SAR backscatter at L- and P-bands and forest biomass is investigated using data acquired within the BioSAR-I campaign in southern Sweden during 2007. Methods for estimating biomass on stand level using these data are developed and evaluated, and the results for the two frequency bands are compared. For L-band data, the best results were obtained using HV-polarized backscatter only, giving estimation errors in terms of root mean square errors (RMSE) between 31% and 46% of the mean biomass for stands with biomass ranging from 10 to 290 t/ha, and an (adjusted) coefficient of determination (R2) between 0.4 and 0.6. For P-band data, the results are better than for L-band. Models using HV- or HH-polarized P-band backscatter give similar results, as does a model including both HV and HH. The RMSEs were between 18 and 27%, and the R2 values were between 0.7 and 0.8.Highlights► L- and P-band backscatter data from a hemi-boreal forest were investigated. ► The P- band data was found to be stable over a period of 2 months. ► Regression models for biomass estimation were developed. ► P-band was found to give lower biomass estimation errors than L-band. ► Using P-band HV and/or HH data, the RMSE was limited to 18-27% of the mean biomass.
The radar backscattering from extraordinary poplar stands, where
all the trunks are leaning towards the same direction, has been
modelled. Data acquired with the JPL/AIRSAR during the MAESTRO-1
campaign have been used. At C- and L-band the backscattering is not
influenced by the leaning of the trunks. At P-band the backscattering
decreases at HH polarisation particularly. This can be predicted
theoretically and is explained by distortion of the trunk-ground
dihedral reflection. A similar effect is obtained when topography
affects the backscattering
Focusing on Siberia, the feasibility of using Advanced Land Observing Satellite Phased Arrayed L-band Synthetic Aperture Radar (PALSAR) summer-intensity and winter-coherence images for large-area forest monitoring was investigated. Fine beam dual horizontal/horizontal and horizontal/vertical (polarization) intensity strip images were acquired during the summer of 2007. The processing consisted of radiometric calibration, orthorectification, and topographic normalization. The coherence was estimated from interferometric pairs with 46-day repeat-pass intervals. The pairs were acquired during the winters of 2006/2007 and 2007/2008. During both winters, suitable weather conditions that allow for low temporal decorrelation had been reported. By using PALSAR intensities and winter-coherence data, areas of forest and nonforest were separated. By combining both data types, a minimal overlap of the class signatures was observed, even though the analysis was conducted at the pixel level and no speckle filter was applied. The study concludes that the operational delineation of forest cover using PALSAR data is feasible. By applying a segmentation-based classification, an accuracy of 93% was obtained.
In this paper the difference between digital elevation models,
derived from P and X bands interferometric data, is used as a main
information to identify land cover classes. The radar data used in this
work were collected on September of 2000 over Tapajos National Forest,
which is a region of Brazilian Amazon, Para State. The SAR images were
acquired from an airborne polarimetric system, AeS-1, that could provide
P and X bands interferometric data. During the radar mission ground
survey was carried out, and the georeferenced information about the
forest typology were acquired, and used as a support for the thematic
identification and calibration of the remoted sensing data. The X-band
DEM was generated using one-pass interferometric data and the P-band DEM
was generated using two-pass interferometric data. The grid of the DEMs
has a spatial resolution of 2.5 meters. Images from P and X bands and
coherence maps were also used in order to improve the classification.
Supervised and unsupervised classifications techniques are used and
their results are shown
This is the first paper in a two-part sequence that evaluates the microwave dielectric behavior of soil-water mixtures as a function of water content, temperature, and soil textural composition. Part I presents the results of dielectric constant measurements conducted for five different soil types at frequencies between 1.4 and 18 GHz. Soil texture is shown to have an effect on dielectric behavior over the entire frequency range and is most pronounced at frequencies below 5 GHz. In addition, the dielectric properties of frozen soils suggest that a fraction of the soil water component remains liquid even at temperatures of -24° C. The dielectric data as measured at room temperature are summarized at each frequency by polynomial expressions dependent upon both the volumetric moisture content m and the percentage of sand and clay contained in the soil; separate polynomial expressions are given for the real and imaginary parts of the dielectric constant. In Part II, two dielectric mixing models will be presented to account for the observed behavior: 1) a semiempirical refractive mixing model that accurately describes the data and requires only volumetric moisture and soil texture as inputs, and 2) a theoretical four-component mixing model that explicitly accounts for the presence of bound water.
The Michigan Microwave Canopy Scattering model (MIMICS) is based on a first-order solution of the radiative-transfer equation for a tree canopy comprising a crown layer, a trunk layer and a rough-surface ground boundary. The crown layer is modelled in terms of distributions of dielectric cylinders (representing needles and/or branches) and discs (representing leaves), and the trunks are treated as dielectric cylinders of uniform diameter. This report describes MIMICS I, which pertains to tree canopies with horizontally continuous (closed) crowns. The model, which is intended for use in the 0·5-10GHz region at angles greater than 10 from normal incidence, is formulated in terms of a 4 × 4 Stokes-like transformation matrix from which the backscattering coefficient can be computed for any transmit/receive polarization configuration.
A microwave scattering model was developed for layered vegetation
based on an iterative solution of the radiative transfer equation up to
the second order to account for multiple scattering within the canopy
and between the ground and the canopy. The model is designed to operate
over a wide frequency range for both deciduous and coniferous forest and
to account for the branch size distribution, leaf orientation
distribution, and branch orientation distribution for each size. The
canopy is modeled as a two-layered medium above a rough interface. The
upper layer is the crown, containing leaves, stems, and branches. The
lower layer is the trunk region, modeled as randomly positioned
cylinders with a preferred orientation distribution above an irregular
soil surface. Comparisons of results obtained using this model with
measurements from deciduous and coniferous forests show good agreement
at several frequencies for both like and cross polarizations
It is shown that the assumptions about the backscatter and the
polarimetric radar system for the two approaches described in J.D.
Klein's (1992) and J.J. van Zyl's (1990) papers are equivalent. It is
demonstrated that, to first order in the radar system crosstalk (i.e.
neglecting terms of second order and above), an exact solution to the
Stokes matrix format data calibration problem exists. It is shown that
van Zyl's approach can give this first-order solution for appropriately
symmetrized polarimetric radar data. Then it is shown how, if the data
are properly symmetrized, van Zyl's approach can be used to calibrate
both scattering matrix and Stokes matrix format data. These conclusions
should be generally applicable to polarimetric imaging radar system data
The authors present the results of an experiment defined to
demonstrate the use of radar to retrieve forest biomass. The SAR data
were acquired by the NASA/JPL SAR over the Landes pine forest during the
1989 MAESTRO-1 campaign. The SAR data, after calibration, were analyzed
together with ground data collected on forest stands from a young stage
(eight years) to a mature stage (46 years). The dynamic range of the
radar backscatter intensity from forest was found to be greatest at
P -band and decreased with increasing frequencies.
Cross-polarized backscatter intensity yielded the best sensitivities to
variations of forest biomass. L -band data confirmed past
results on good correlation with forest parameters. The most striking
observation was the strong correlation of P -band backscatter
intensity to forest biomass
Two independent experimental efforts have examined the dependence
of radar backscatter on above-ground biomass of monospecie conifer
forests using polarimetric airborne SAR data at P -, L -
and C -bands. Plantations of maritime pines near Landes, France,
range in age from 8 to 46 years with above-ground biomass between 5 and
105 tons/ha. Loblolly pine stands established on abandoned agricultural
fields near Duke, NC, range in age from 4 to 90 years and extend the
range of above-ground biomass to 560 tons/ha for the older stands. These
two experimental forests are largely complementary with respect to
biomass. Radar backscatter is found to increase approximately linearly
with increasing biomass until it saturates at a biomass level that
depends on the radar frequency. The biomass saturation level is about
200 tons/ha at P -band and 100 tons/ha at L -band, and
the C -band backscattering coefficient shows much less
sensitivity to total above-ground biomass
The peak and integral methods for radiometric calibration of a
synthetic aperture radar (SAR) using reference point targets are
analyzed. Both calibration methods are shown to be unbiased, but the
peak method requires knowledge of the equivalent rectangle system
resolution which is sensitive to system focus. Exact expressions for the
RMS errors of both methods are derived. It is shown that the RMS error
resulting from the peak method is always smaller than or equal to that
from the integral method for a well-focused system. However, for robust
radiometric calibration of SAR, or when nonlinear phase errors are
present, the integral method is recommended, because it does not require
detailed knowledge of the impulse response and the resulting RMS error
is not dependent on system focus
Procedures for the collection and compilation of forest ground data in microwave experiments
- P N Dobson
- F T Ulaby
- Le Toan
- T Beaudoin
- A Kasischke
- E S Christensen
References CHURCHILL, P. N., 1989, Procedures for the collection and compilation of forest ground data in microwave experiments, JRC Ispra Technical Note, July. DOBSON, c., ULABY, F. T., LE TOAN, T., BEAUDOIN, A., KASISCHKE, E. S., and CHRISTENSEN, N., 1992, Dependence of radar backscatter on coniferous forest biomass, I.E.E.£. Transactions on Geoscience and Remote Sensing, 30, 412-415.
Computer aided forest stand delineation and inventory based on satellite remote sensing
- Hagner
HAGNER, 0., 1990, Computer aided forest stand delineation and inventory based on satellite remote sensing, Proceedings of the SNS/IUFRO Workshop. Umeb, 26-28 February, Report no. 4, (Umea: Swedish University of Agricultural Science), pp. 94-105.
A study on the quality of compartmentwise forest data acquired by subjective inventory methods , Report No. 24 , Department of Biometry and Forest Management
- G Stahl
STAHL, G., 1992, A study on the quality of compartmentwise forest data acquired by subjective inventory methods, Report No. 24, Department of Biometry and Forest Management, Swedish University of Agricultural Science, Umea, Sweden. MAESTRO J airborne polarimetric SAR campaign ULARY, F. T., and EL-RAYEs, M. A., 1987, Microwave dielectric spectrum of vegetation -Part
Radar remote sensing for applications in forestry. A literature review for investigators and potential users of SAR data in Canada, Canada Centre for Remote Sensing, Technology Transfer Program/Applications Division
- D Werle
WERLE, D., 1989, Radar remote sensing for applications in forestry. A literature review for investigators and potential users of SAR data in Canada, Canada Centre for Remote Sensing, Technology Transfer Program/Applications Division, Ottawa, Ontario, Canada. Downloaded by [University of Bristol] at 03:45 04 March 2015