Gustaf Sandberg

Detection of forest changes using ALOS PALSAR satellite images

A controlled experiment has been performed to quantify the ability to detect clear-cuts using ALOS PALSAR data. The experiment consisted of 8 old spruce dominated stands, each with a size of about 1.5 ha, located at a test site in southern Sweden. Four of the stands were clear-felled and the remaining stands were left untreated for reference. A time series of PALSAR images was acquired prior to, during, and after treatment, including 7 fine beam single polarization (FBS, look angle 34.3deg, HH-polarization) SAR images. The results clearly show that the clear-felled stands could be separated from the reference stands. The drop in backscattering coefficient between the reference and the clear-felled stands was on average 2.1 dB. This implies that ALOS PALSAR data potentially can be used for large-scale mapping of changes in forest cover.

A Physical-Optics Model for Double-Bounce Scattering From Tree Stems Standing on an Undulating Ground Surface

In this paper, a model for prediction of radar backscatter from coniferous forests in the VHF and UHF band is proposed. The model includes the double-bounce scattering originating from vertical stems standing on an undulating ground surface and is based on a physical-optics approach. The model can be used to assess the importance of ground topography in synthetic aperture radar (SAR) imagery of forests, and it is applicable to SAR systems using horizontally transmit and receive polarization (HH). The model was validated against data from the airborne SAR systems CARABAS-II and LORA. Precision measurements of ground topography and forest characterization at a single tree level were used as model input to simulate SAR images. The simulated images were compared to radar data in the frequency bands 22-82 and 225- 470 MHz, and it was found that the model could predict much of the variation in backscatter observed in images (R2 = 0.44 and 0.65 at best, for the lower and higher frequency band, respectively), which should be compared to R2 = 0.1 if the same model, but assuming a flat ground, was used. The results thus indicate that ground topography must be considered when predicting the variations in backscatter in the SAR images studied. The model did, however, fail to predict the absolute values of the backscattered intensity. The reason for the discrepancy is believed to be the value chosen for stem dielectric constant and unmodeled effects due to wave attenuation, tilting stems, and small-scale surface roughness.

Topographic correction for biomass retrieval from P-band SAR data in boreal forests

The influence of the ground slope on radar backscatter has been proven to be greater for lower radar frequencies due to deeper canopy penetration. In this study, multiple heading, P-band SAR data of boreal forest in Sweden was used to find a model for topographic correction for improved biomass retrieval. Eleven models were tested and the best model was selected. The selected model was then used for biomass retrieval. Even by means of the most simplified approach, forest biomass could be established with a root-mean-square error of approximately 50 t/ha for HV and 66 t/ha for HH.

Measurements of Faraday Rotation Using Polarimetric PALSAR Images

For spaceborne synthetic aperture radar (SAR) systems operating at L-band frequencies or lower, the ionosphere may have a significant impact on the SAR images. The largest effect at L-band is caused by Faraday rotation (FR). Several studies have modeled the effect of FR and/or devised models to measure and correct FR. With the launch of the fully polarimetric L-band system Phased Array-type L-band SAR (PALSAR), it has become possible to test both models and measurement techniques on real SAR data. In this letter, the quality of calibrated polarimetric PALSAR data is assessed, and FR is measured. It is found that residual crosstalk and channel imbalance are small in the PALSAR data. Two methods are used to measure FR, the first using in-scene distributed targets and the second using large trihedrals. The two methods show very good agreement. The measurements are compared with values of the total electron content using a linear model. It is found that the model and measurements are in good agreement, with a root-mean-square error of 0.3deg or 15% of the mean FR angle.

Biomass retrieval algorithm based on P-band biosar experiments of boreal forest

A new biomass retrieval algorithm based on P-band multi-polarization backscatter has been developed and evaluated based on SAR and ground data over boreal forest. SAR data collections were conducted on three dates at a test site in southern Sweden (Remningstorp, biomass < 300 tons/ha; late winter to early summer 2007) and on a single date at a test site in northern Sweden (Krycklan, biomass < 200 tons/ha; fall 2008). The retrieval algorithm is a multiple linear regression model including the HV-polarized backscatter coefficient, the VV/HH backscatter ratio and the ground slope. Regression coefficients were determined from Krycklan data followed by algorithm evaluation using Remningstorp data. The results from the latter show that RMS errors vary in the range 29–42 tons/ha depending on date and stand type. The new algorithm is also compared with alternative algorithms and found to give significantly better performance. The developed model is a significant step towards an algorithm which gives consistent results across multiple sites and dates, i.e. when forest structure, topography and moisture conditions is expected to vary.

ALOS PALSAR Calibration and Validation Results from Sweden

In 2006 calibration activities for ALOS PALSAR were conducted in Sweden. Four five-metre trihedral corner reflectors and three smaller dihedral reflectors were deployed and operated during eight months. 23 PALSAR scenes were acquired over the calibration site allowing an evaluation of the quality and temporal stability of the data. Results show that the co-polarized data have been stable during the whole calibration period with variations in the trihedral responses lower than 0.7 dB. The measured resolution in azimuth was 4.4 m and in slant range 4.7 m for single polarization images and 9.5 m for polarimetric data. For the cross-polarized data large variations in the dihedral responses were found. It is assumed that this is caused by a larger sensitivity to pointing errors. For the polarimetric data, estimation of Faraday rotation gave values ranging from 0.1deg to 3deg.

Statistical Analysis of VHF-Band Tree Backscattering Using Forest Ground Truth Data and PO Scattering Model

This paper analyzes the statistical properties of the very high frequency (VHF)-band radar backscattering from coniferous trees by incorporating forest ground truth data into a physical-optics (PO) model that assumes horizontally transmit and receive polarizations and dominant double-bounce scattering from vertical stems standing on an undulating ground surface. The analysis shows that a statistically adequate model for the tree backscattering amplitude can be presented as a mixture of generalized gamma or lognormal distribution, and the mixture model can be reduced to a single density model if the trees with trunk volumes exceeding an appropriate threshold are to be taken into account. The generalized gamma density is shown to provide an appreciably better fit to the exceedance functions associated with the PO model data than that for the lognormal density. The results can be used to design statistically adequate models of forest clutter for VHF synthetic aperture radar systems.

Comparison of L- and P-band biomass retrievals based on backscatter from the BioSAR campaign

With the continued threat of global warming, the need to obtain consistent and accurate measurements of the carbon stored in forests is strong. L- and P-band SAR backscatter data have shown to be sensitive to forest biomass, which in turn is coupled to the stored carbon. In this paper a biomass retrieval method is developed for L- and P-band using data from the BioSAR campaign conducted in Sweden during the spring 2007 over hemi-boreal forest. The results show that the use of L-band data gives an underestimation of biomass for stands with high biomass; while for P-band no such underestimation is seen. RMSEs are found to be 30–40% of the mean biomass for L-band and about 25% for P-band for stands with biomass ranging from 10 to 290 tons/ha.

Mapping of wind-thrown forests using VHF/UHF SAR images

SAR images from the Swedish airborne CARABAS-II and LORA systems have been visually analyzed over simulated wind-thrown forest at both single tree and stand level. In ideal conditions, the results show that LORA is more accurate than CARABAS-II at detecting wind-thrown trees, regardless of tree size and direction of the fallen trees relative to flight heading. Furthermore, the visible single trees in the LORA images appeared more distinct than in the CARABAS-II images, which could be explained by the high resolution in the LORA images. Based on visual interpretation, it is likely that the detection of wind-thrown forests could be improved using VHF/UHF SAR images acquired both prior to and after a storm event.

Impact and modeling of topographic effects on P-band SAR backscatter from boreal forests

P-band SAR backscatter has been proven to be useful for forest biomass prediction. However, there is a need for further studies on effects of topography on P-band backscatter. In this paper, two prediction models for backscatter are evaluated, one using only biomass as predictor and one which also includes topographic corrections. Data from the BioSAR 2007 and BioSAR 2008 campaigns are used to evaluate the models. A multi-scale error model which is able to handle data from several imaging directions is used. For HH, the slope correction on stand level used in this paper is unable to correct for topographic effects. This is consistent with previous results that within stand topographic variability has a significant impact on HH P-band backscatter. For HV and VV, the model which considers topography gives lower prediction errors than the model which does not include topography. Moreover, for these polarizations topographic the correction strongly reduce the variability in backscatter measurements between imaging directions for stands with ground slopes larger than about 5 degrees.