Interferometric Water Cloud Model Research Articles

Sentinel-1 Coherence for Mapping Above-Ground Biomass in Semiarid Forest Areas

The interferometric coherence of repeat-pass C-band synthetic aperture radar observations has shown potential for mapping forest above-ground biomass (AGB) when images were acquired with short temporal repeat intervals and under stable environmental conditions. Since 2014, Sentinel-1 collects repeated interferometric image pairs at global scale with repeat intervals of 6 or 12 days. For a semiarid forest site in California, we evaluated one year of Sentinel-1 6- and 12-day (60 and 59 images, respectively) repeat-pass coherence images with the objective of retrieving AGB. The retrieval was performed by inverting the semiempirical Interferometric Water Cloud Model. The retrieval based on individual coherence images was most accurate (~50% relative root-mean-square difference with respect to a Lidar-derived map) when images were acquired in extended periods of dry conditions. The retrieval accuracy improved by approximately 15% units when combining multiple coherence images in the retrieval. Finally, the accuracy of the retrieval with multitemporal 6- or 12-day repeat-pass coherence images differed by only 2% units. The results of this study advocate consideration of Sentinel-1 repeat-pass coherence in AGB mapping efforts in semiarid forest areas.

Aboveground biomass estimates of tropical mangrove forest using Sentinel-1 SAR coherence data - The superiority of deep learning over a semi-empirical model

The availability of advanced Machine Learning algorithms has made the estimation process of biophysical parameters more efficient. However, the efficiency of those methods seldom compared with the efficiency of already established semi-empirical procedures. Aboveground biomass (AGB) of mangrove forests is a crucial biophysical parameter as it is positively correlated to the carbon stocks and fluxes. The free availability of Sentinel-1 C-band SAR data and machine learning algorithms hold promises in estimating AGB of tropical mangrove forests. We reported high AGB (70 t/ha to 666 t/ha) using 185 field quadrats of 0.04ha each from Bhitarkanika Wildlife Sanctuary, located on the eastern Indian coast that could be attributed to species composition. The AGB maps generated using Interferometric Water Cloud Model (IWCM) and Deep Learning models were different from each other as they rely on different variables. IWCM was more dependent, especially on ground and vegetation components of coherence, while canopy height acted as the most crucial variable in the Deep Learning model. However, the negligible variations in Deep Learning-based AGB maps can be attributed to interpreting the importance of coherence and VH backscatter. Due to low canopy penetration power of C-band SAR, high temporal decorrelation resulting from longer time gap between interferometric image pairs, and high spatial heterogeneity of mangrove forests, IWCM found as an unsuitable method for AGB estimation. Interestingly, a Deep Learning algorithm could translate the exact relationship between predictor variables and mangrove AGB in Bhitarkanika Wildlife Sanctuary. The AGB estimation studies in mangrove forests using Sentinel data should focus more on using machine learning algorithms like Deep Learning rather than semi-empirical models.

Boreal Forest Properties From TanDEM-X Data Using Interferometric Water Cloud Model and Implications for a Bistatic C-Band Mission

Data from TanDEM-X in single-pass and bistatic interferometric mode together with the interferometric water cloud model (IWCM) can provide estimates of forest height and stem volume (or the related above-ground biomass) of boreal forests with high accuracy. We summarize results from two boreal test sites using two approaches, i.e., 1) based on model calibration using reference insitu stands, and 2) based on minimization of a cost function. Both approaches are based on inversion of IWCM, which models the complex coherence and backscattering coefficient of a homogeneous forest layer, which includes gaps where free-space wave propagation is assumed. A digital terrain model of the ground is also needed. IWCM is used to estimate forest height or stem volume, since the two variables are assumed to be related through an allometric equation. A relationship between the fractional area of gaps, the area-fill, and stem volume is also required to enable model inversion. The accuracy of the stem volume estimate in the two sites varies between 16% and 21% for height of ambiguity <100 m. The results clearly show the importance of using summer-time acquisitions. Based on the TanDEM-X results at X-band, C-band data from the ERS-1/ERS-2 tandem mission are revisited to investigate the potential of a future bistatic C-band interferometric mission. Out of nine ERS-1/ERS-2 pairs, only one pair was found to be acquired at summer temperatures, without precipitation and with high coherence. A simulated bistatic phase height is shown to give approximately the same sensitivity to stem volume as TanDEM-X.

On the Sensitivity of TanDEM-X-Observations to Boreal Forest Structure

The structure of forests is important to observe for understanding coupling to global dynamics of ecosystems, biodiversity, and management aspects. In this paper, the sensitivity of X-band to boreal forest stem volume and to vertical and horizontal structure in the form of forest height and horizontal vegetation density is studied using TanDEM-X satellite observations from two study sites in Sweden: Remningstorp and Krycklan. The forest was analyzed with the Interferometric Water Cloud Model (IWCM), without the use of local data for model training, and compared with measurements by Airborne Lidar Scanning (ALS). On one hand, a large number of stands were studied, and in addition, plots with different types of changes between 2010 and 2014 were also studied. It is shown that the TanDEM-X phase height is, under certain conditions, equal to the product of the ALS quantities for height and density. Therefore, the sensitivity of phase height to relative changes in height and density is the same. For stands with a phase height &gt;5 m we obtained an root-mean-square error, RMSE, of 8% and 10% for tree height in Remningstorp and Krycklan, respectively, and for vegetation density an RMSE of 13% for both. Furthermore, we obtained an RMSE of 17% for estimation of above ground biomass at stand level in Remningstorp and in Krycklan. The forest changes estimated with TanDEM-X/IWCM and ALS are small for all plots except clear cuts but show similar trends. Plots without forest management changes show a mean estimated height growth of 2.7% with TanDEM-X/IWCM versus 2.1% with ALS and a biomass growth of 4.3% versus 4.2% per year. The agreement between the estimates from TanDEM-X/IWCM and ALS is in general good, except for stands with low phase height.

Forest stem volume estimation using C-band interferometric SAR coherence data of the ERS-1 mission 3-days repeat-interval phase

Interferometric synthetic aperture radar (InSAR) coherence datasets from the 3-days phase of the European Remote Sensing Satellite (ERS-1) mission during the winter months of 1992 have been analyzed to assess the capability to retrieve forest stem volume in boreal forests. For three test sites in Sweden and Finland, coherence decreased for increasing repeat-pass interval from 3- to 12-days due to increasing temporal decorrelation. Overall strong decorrelation was observed for images acquired under unfrozen conditions or thawing conditions. Coherence was best preserved under stable frozen conditions and, in addition, short perpendicular component of the interferometric baseline (i.e., <300 m). Stem volume was estimated with the Interferometric Water Cloud Model (IWCM) and multi-temporal combination of single-image estimates. For the test site of Tuusula, Finland, data from eight 3-days image pairs yielded a relative RMSE of 28.2% for stem volumes ranging between 3 and 482 m3/ha. For the 6-days repeat-pass interval, the relative RMSE was only slightly worse (32.4%). For longer repeat-pass intervals, the relative RMSE was above 50%. This study suggests that C-band short-term coherence has potential to support retrieval of forest biomass but the prospects of Sentinel-1 coherence to estimate biomass appear to be limited to the 6-days coherence in areas characterized by stable frozen, or likewise stable dry, conditions.

Biomass Growth from Multi-Temporal TanDEM-X Interferometric Synthetic Aperture Radar Observations of a Boreal Forest Site

Forest growth estimation is important in forest research and forest management, but complex to analyze in diverse forest stands. Twelve summertime TanDEM-X acquisitions from the boreal test site, Krycklan, in Sweden, with a known digital terrain model, DTM, have been used to study phase height and aboveground biomass change over 3.2 years based on the Interferometric Water Cloud Model, IWCM. The maximum phase height rate was determined to 0.29 m/yr, while the mean phase height rate was 0.16 m/yr. The corresponding maximum growth rate of the aboveground dry biomass, AGB, was 4.0 Mg/ha/yr with a mean rate of 1.9 Mg/ha/yr for 27 stands, varying from 23 to 183 Mg/ha. The highest relative AGB growth was found for young stands and high growth rates up to an age of 150 years. Growth rate differences relative a simplified model assuming AGB to be proportional to the phase height were studied, and the possibility to avoid a DTM was discussed. Effects of tree species, thinning, and clear cutting were evaluated. Verifications using in situ data from 2008 and a different in situ dataset combined with airborne laser scanning data from 2015 have been discussed. It was concluded that the use of multi-temporal TanDEM-X interferometric synthetic aperture radar observations with AGB estimates of each individual observation can be an important method to derive growth rates in boreal forests.

Estimation of Boreal Forest Properties From TanDEM-X Data Using Inversion of the Interferometric Water Cloud Model

In this letter, the interferometric water cloud model (IWCM) is fit to 87 VV-polarized TanDEM-X acquisitions made between June 2011 and August 2014 over a boreal forest in Krycklan, northern Sweden, using a new method based on nonlinear least-squares optimization. A high-resolution digital terrain model is used as ground reference during interferometric synthetic-aperture radar (InSAR) processing and 26 stands with areas 1.5–22 ha and unaltered during the study period are studied. The dependence of biomass estimation performance, ground and vegetation backscatter coefficients ( $\sigma ^{0}_{\mathrm {gr}}$ and $\sigma ^{0}_{\mathrm {veg}}$ ), canopy attenuation ( $\alpha$ ), and zero-biomass coherence ( $\gamma _{0}$ ) on selected system and environmental parameters is studied. High correlation between the estimated biomass and reference biomass derived from in situ measurements is observed for all 87 acquisitions ( $r$ between 0.81 and 0.93), while the root-mean-square difference is between 18% and 32% for all 43 acquisitions made in snow-free conditions and with heights-of-ambiguity (HOAs) between 36 and 150 m. Significant biomass estimation bias is observed for HOAs above 150 m and for some acquisitions over snow-covered forest. It is also observed that $\sigma ^{0}_{\mathrm {gr}}$ and $\sigma ^{0}_{\mathrm {veg}}$ are the largest for temperatures below 0 °C and with significant snow cover. For temperatures above 0 °C, $\sigma ^{0}_{\mathrm {gr}}$ appears independent of temperature, while $\sigma ^{0}_{\mathrm {veg}}$ shows a tendency to increase with temperature. Moreover, $\gamma _{0}$ decreases from just below 1 for HOAs around 40 m to around 0.8 for HOAs above 150 m.

Biomass estimation in a boreal forest from TanDEM-X data, lidar DTM, and the interferometric water cloud model

The semi-empirical Interferometric Water Cloud Model, IWCM, is used to estimate above ground forest biomass, AGB, in northern Sweden, Krycklan (64°N 20°E). The results are based on separate analysis of 14 TanDEM-X acquisitions from 2011 to 2014 and a Lidar digital terrain model (DTM). 29 stands covering 272ha and with AGB<183Mg/ha, and 619 stands with area>1ha covering 3166ha and with AGB<291Mg/ha have been analyzed. In situ and airborne lidar scanning, ALS, data from the BioSAR 2008 experiment are used as reference. AGB and forest height are estimated using a new optimization method for determining IWCM parameters. Allometric equations are used to describe the inter-dependency between forest height, biomass, area-fill factor, and stem volume. No local training data from the investigated area are used to determine model parameters. For the 29 stands, the relative RMSE for biomass estimated using the proposed method varied between 15.8% and 21.2% (r2 between 0.82 and 0.88) and between 9.9% and 16.0% for height (r2 between 0.84 and 0.89). Dependence of model parameters on temperature and precipitation as well as height of ambiguity are investigated. A method based on look-up table for biomass estimation from phase height is proposed. The method is used over an area of 68km2 for one TanDEM-X acquisition from 2011-06-04 and the results are compared with an ALS biomass map from August 2008. Good agreement is observed, as well as high potential for clear-cut detection.

Polarimetric SAR Interferometry based modeling for tree height and aboveground biomass retrieval in a tropical deciduous forest

The regulation of the carbon cycle is a critical ecosystem service provided by forests globally. It is, therefore, necessary to have robust techniques for speedy assessment of forest biophysical parameters at the landscape level. It is arduous and time taking to monitor the status of vast forest landscapes using traditional field methods. Remote sensing and GIS techniques are efficient tools that can monitor the health of forests regularly. Biomass estimation is a key parameter in the assessment of forest health. Polarimetric SAR (PolSAR) remote sensing has already shown its potential for forest biophysical parameter retrieval. The current research work focuses on the retrieval of forest biophysical parameters of tropical deciduous forest, using fully polarimetric spaceborne C-band data with Polarimetric SAR Interferometry (PolInSAR) techniques. PolSAR based Interferometric Water Cloud Model (IWCM) has been used to estimate aboveground biomass (AGB). Input parameters to the IWCM have been extracted from the decomposition modeling of SAR data as well as PolInSAR coherence estimation. The technique of forest tree height retrieval utilized PolInSAR coherence based modeling approach. Two techniques – Coherence Amplitude Inversion (CAI) and Three Stage Inversion (TSI) – for forest height estimation are discussed, compared and validated. These techniques allow estimation of forest stand height and true ground topography. The accuracy of the forest height estimated is assessed using ground-based measurements. PolInSAR based forest height models showed enervation in the identification of forest vegetation and as a result height values were obtained in river channels and plain areas. Overestimation in forest height was also noticed at several patches of the forest. To overcome this problem, coherence and backscatter based threshold technique is introduced for forest area identification and accurate height estimation in non-forested regions. IWCM based modeling for forest AGB retrieval showed R2 value of 0.5, RMSE of 62.73 (t ha−1) and a percent accuracy of 51%. TSI based PolInSAR inversion modeling showed the most accurate result for forest height estimation. The correlation between the field measured forest height and the estimated tree height using TSI technique is 62% with an average accuracy of 91.56% and RMSE of 2.28m. The study suggested that PolInSAR coherence based modeling approach has significant potential for retrieval of forest biophysical parameters.

On the Estimation of Boreal Forest Biomass From TanDEM-X Data Without Training Samples

Boreal forests play an important part in the climate system, and estimates of the biomass are important also from an economic point of view. In this letter, forest aboveground biomass is estimated from bistatic TanDEM-X data, a Lidar digital elevation model (DEM), and the interferometric water cloud model, without using training samples to calibrate the model. The forest was characterized by allometric relations for area fill (vegetation fraction) and height versus stem volume, and stem volume versus biomass. Biomass was estimated for 202 forest stands at least 1 ha large at the forest test site of Remningstorp, Sweden, from 18 bistatic TanDEM-X acquisitions with a relative root-mean-square error (RMSE) of 16%–32%. TanDEM-X acquisitions with a height of ambiguity around 80 m resulted in the best results. A multitemporal combination resulted in a relative RMSE of 17%. This result is comparable with the retrieval error obtained in a previous study when training the model using a set of known forest stands.

Model-Based Biomass Estimation of a Hemi-Boreal Forest from Multitemporal TanDEM-X Acquisitions

Above-ground forest biomass is a significant variable in the terrestrial carbon budget, but is still estimated with relatively large uncertainty. Remote sensing methods can improve the characterization of the spatial distribution and estimation accuracy of biomass; in this respect, it is important to examine the potential offered by new sensors. To assess the contribution of the TanDEM-X mission, eighteen interferometric Synthetic Aperture Radar (SAR) image pairs acquired over the hemi-boreal test site of Remningstorp in Sweden were investigated. Three models were used for interpretation of TanDEM-X signatures and above-ground biomass retrieval: Interferometric Water Cloud Model (IWCM), Random Volume over Ground (RVoG) model, and a simple model based on penetration depth (PD). All use an allometric expression to relate above-ground biomass to forest height measured by TanDEM-X. The retrieval was assessed on 201 forest stands with a minimum size of 1 ha, and ranging from 6 to 267 Mg/ha (mean biomass of 105 Mg/ha) equally divided into a model training dataset and a validation test dataset. Biomass retrieved using the IWCM resulted in a Root Mean Square Error (RMSE) between 17% and 33%, depending on acquisition date and image acquisition geometry (angle of incidence, interferometric baseline, and orbit type). The RMSE in the case of the RVoG and the PD models were slightly higher. A multitemporal estimate of the above-ground biomass using all eighteen acquisitions resulted in an RMSE of 16% with R2 = 0.93. These results prove the capability of TanDEM-X interferometric data to estimate forest aboveground biomass in the boreal zone.

Aboveground biomass estimation of tropical forest from Envisat advanced synthetic aperture radar data using modeling approach

100 t/ha), remains a challenging task for the researchers worldwide. The retrieval of AGB over a tropical forest area in India using Envisat advanced synthetic aperture radar C-band backscatter, interferometric synthetic aperture radar (InSAR) coherence and semi-empirical models viz., water cloud model (WCM) and interferometric water cloud model (IWCM), is studied. In process, the model parameters, i.e., backscatter from vegetation and ground, two-way tree transmissivity, and coherence from vegetation and ground were retrieved. The model training procedure to retrieve the model parameters consisted of an iterative regression of WCM and IWCM. High AGB accuracy (R2=0.73) with low root mean square error (RMSE=53.76 t/ha) was achieved through multidate weighted averaging using RMSE-based weighting coefficients and WCM. Multidate data and InSAR coherence images showed better results (R2=0.90, RMSE=35.92 t/ha) compared to individual coherence images. The InSAR coherence was found to be better for AGB retrieval than SAR backscatter as the former did not saturate for high AGB values.

Signatures of ERS–Envisat Interferometric SAR Coherence and Phase of Short Vegetation: An Analysis in the Case of Maize Fields

Interferometric observations between the European Remote Sensing, ERS-2, synthetic aperture radar (SAR) and the Envisat Advanced SAR (ASAR) are unique since they are characterized by a short repeat-pass interval (28 min) and a perpendicular baseline of approximately 2 km. In vegetated areas, this configuration should preserve from strong temporal decorrelation and enhance the sensitivity of coherence and SAR interferometric (InSAR) phase to volumes with small heights. This assumption could be tested with the data acquired during the dedicated ERS-Envisat Tandem mission on October 15, 2007, over the Seeland region, Switzerland. Five maize fields and one sunflower field presented lower coherence and offsets of the interferometric phase, i.e. height, with respect to neighboring bare fields. To gain understanding on the interferometric signatures, the interferometric water cloud model was used to simulate coherence and InSAR height for the maize fields. Both the coherence and the InSAR height present clear dependence upon vegetation height and exhibit strong consistency. Simulations showed that the modeled coherence and InSAR height are most sensitive to the two-way attenuation and the temporal coherence of the vegetation. The best correspondence between the observed and modeled InSAR parameters was obtained with two-way attenuation values between 2 and 4 dB/m (corresponding to an extinction between 1 and 2 dB/m) and high temporal coherence of the vegetation (above 0.6), with this being due to the very stable conditions of the weather during the 28-min interval between image acquisitions.

Multitemporal repeat pass SAR interferometry of boreal forests

Multitemporal interferometric European Remote Sensing 1 and 2 satellite tandem pairs from a forest test site in Finland are examined in order to determine the stem volume retrieval accuracy. A form of multitemporal filtering is introduced to investigate what forest stands show a multitemporal consistency in coherence. It is found that a large stand size is a major factor to obtain accurate retrievals. The effect of heterogeneity of forest stands is also discussed. Based on the stands showing highest multitemporal consistency different models for scattering and coherence are compared. The interferometric water cloud model is chosen for stem volume retrieval. The variation of the model parameters with meteorological parameters is investigated and the results illustrate that the best imaging conditions are obtained for subzero temperatures and windy conditions. It is shown that for the 20 stands showing highest multitemporal consistency the stem volume can be retrieved with a relative error of 21%, deteriorating when the number of testing stands is increased, e.g., for 80 stands the error is 48%. For 37 large forest stands representing 48% of the investigated area the relative stem volume error is 26%. With experience from another site in Sweden we may conclude that the error level for a multitemporal interferometric synthetic aperture radar evaluation of stem volume for large forest stands (>2 ha) in a well managed and homogeneous boreal forest may be expected to be in the 15% to 25% range, deteriorating for small and heterogeneous stands and for images acquired under nonwinter conditions.

Tree height influence on ERS interferometric phase in boreal forest

The European Remote Sensing 1/2 (ERS-1/2) "tandem" coherence has been shown to provide estimates of stem volume in boreal forest in agreement with in situ data. Tree height estimation from ERS interferometric phase represents a further step in the investigations concerning the retrieval of biophysical parameters using repeat-pass synthetic aperture radar (SAR) interferometry. At two test sites located in Sweden and Finland, sets of respectively nine and eight ERS "tandem" interferograms were available. Images acquired under stable winter weather conditions and during nighttime were found to be less affected by atmospheric artifacts. Reduction of atmospheric artifacts in interferograms was performed with a phase screen estimated over a dense grid of open areas. Nonetheless, at each test site, only a limited set of pairs was useful for tree height investigations. Under stable winter conditions, the interferometric tree height obtained from an inversion of the differential interferometric phase at stand level was found to be much lower than the true tree height. Spread and uncertainty in the interferometric tree height measurements were caused by phase noise and residual atmospheric artifacts. Using the semiempirical interferometric water cloud model (IWCM), the modeled interferometric tree height was generally in reasonable agreement with the measurements, showing the need of a phase term in interferometric modeling of forests. The inversion of the IWCM for tree height retrieval showed the strong effect of phase noise and atmospheric artifacts on the estimates. Hence, tree height retrieval from ERS repeat-pass SAR interferometry seems to have limited forestry applications. The results also indicate under what conditions the forest influence is small on digital elevation models derived from repeat-pass interferometry.