A systematic method for 3D mapping of mangrove forests based on Shuttle Radar Topography Mission elevation data, ICEsat/GLAS waveforms and field data: Application to Ciénaga Grande de Santa Marta, Colombia

Mangrove forests are found within the intertropical zone and are one of the most biodiverse and productive wetlands on Earth. We focus on the Ciénaga Grande de Santa Marta (CGSM) in Colombia, the largest coastal lagoon–delta ecosystem in the Caribbean area with an extension of 1280 km², where one of the largest mangrove rehabilitation projects in Latin America is currently underway. Extensive man-made hydrological modifications in the region caused hypersaline soil (> 90 g kg^− 1) conditions since the 1960s triggering a large dieback of mangrove wetlands (~ 247 km²). In this paper, we describe a new systematic methodology to measure mangrove height and aboveground biomass by remote sensing. The method is based on SRTM (Shuttle Radar Topography Mission) elevation data, ICEsat/GLAS waveforms (Ice, Cloud, and Land Elevation Satellite/Geoscience Laser Altimeter System) and field data. Since the locations of the ICEsat and field datasets do not coincide, they are used independently to calibrate SRTM elevation and produce a map of mangrove canopy height. We compared height estimation methods based on waveform centroids and the canopy height profile (CHP). Linear relationships between ICEsat height estimates and SRTM elevation were derived. We found the centroid of the canopy waveform contribution (CWC) to be the best height estimator. The field data was used to estimate a SRTM canopy height bias (− 1.3 m) and estimation error (rms = 1.9 m). The relationship was applied to the SRTM elevation data to produce a mangrove canopy height map. Finally, we used field data and published allometric equations to derive an empirical relationship between canopy height and biomass. This relationship was used to scale the mangrove height map and estimate aboveground biomass distribution for the entire CGSM. The mean mangrove canopy height in CGSM is 7.7 m and most of the biomass is concentrated in forests around 9 m in height. Our biomass maps will enable estimation of regeneration rates of mangrove forests under hydrological rehabilitation at large spatial scales over the next decades. They will also be used to assess how highly disturbed mangrove forests respond to increasing sea level rise under current global climate change scenarios.     

Satellite lidar vs. small footprint airborne lidar: Comparing the accuracy of aboveground biomass estimates and forest structure metrics at footprint level

The use of lidar remote sensing for mapping the spatial distribution of canopy characteristics has the potential to allow an accurate and efficient estimation of tree dimensions and canopy structural properties from local to regional and continental scales. The overall goal of this paper was to compare biomass estimates and height metrics obtained by processing GLAS waveform data and spatially coincident discrete-return airborne lidar data over forest conditions in east Texas. Since biomass estimates are derived from waveform height metrics, we also compared ground elevation measurements and canopy parameters. More specific objectives were to compare the following parameters derived from GLAS and airborne lidar: (1) ground elevations; (2) maximum canopy height; (3) average canopy height; (4) percentiles of canopy height; and (5) above ground biomass. We used the elliptical shape of GLAS footprints to extract canopy height metrics and biomass estimates derived from airborne lidar. Results indicated a very strong correlation for terrain elevations between GLAS and airborne lidar, with an r value of 0.98 and a root mean square error of 0.78 m. GLAS height variables were able to explain 80% of the variance associated with the reference biomass derived from airborne lidar, with an RMSE of 37.7 Mg/ha. Most of the models comparing GLAS and airborne lidar height metrics had R-square values above 0.9.     

Estimating aboveground biomass of a mangrove plantation on the Northern coast of Vietnam using machine learning techniques with an integration of ALOS-2 PALSAR-2 and Sentinel-2A data

ABSTRACT

Aboveground biomass (AGB) of mangrove forest plays a crucial role in global carbon cycle by reducing greenhouse gas emissions and mitigating climate change impacts. Monitoring mangrove forests biomass accurately still remains challenging compared to other forest ecosystems. We investigated the usability of machine learning techniques for the estimation of AGB of mangrove plantation at a coastal area of Hai Phong city (Vietnam). The study employed a GIS database and support vector regression (SVR) to build and verify a model of AGB, drawing upon data from a survey in 25 sampling plots and an integration of Advanced Land Observing Satellite-2 Phased Array Type L-band Synthetic Aperture Radar-2 (ALOS-2 PALSAR-2) dual-polarization horizontal transmitting and horizontal receiving (HH) and horizontal transmitting and vertical receiving (HV) and Sentinel-2A multispectral data. The performance of the model was assessed using root mean square error (RMSE), mean absolute error (MAE), coefficient of determination (R²), and leave-one-out cross-validation. Usability of the SVR model was assessed by comparing with four state-of-the-art machine learning techniques, i.e. radial basis function neural networks, multi-layer perceptron neural networks, Gaussian process, and random forest. The SVR model shows a satisfactory result (R² = 0.596, RMSE = 0.187, MAE = 0.123) and outperforms the four machine learning models. The SVR model-estimated AGB ranged between 36.22 and 230.14 Mg ha^?1 (average = 87.67 Mg ha^?1). We conclude that an integration of ALOS-2 PALSAR-2 and Sentinel-2A data used with SVR model can improve the AGB accuracy estimation of mangrove plantations in tropical areas.     

Regional aboveground forest biomass using airborne and spaceborne LiDAR in Québec

Aboveground dry biomass was estimated for the 1.3 M km² forested area south of the treeline in the eastern Canadian province of Québec by combining data from an airborne and spaceborne LiDAR, a Landsat ETM+ land cover map, a Shuttle Radar Topographic Mission (SRTM) digital elevation model, ground inventory plots, and vegetation zone maps. Plot-level biomass was calculated using allometric relationships between tree attributes and biomass. A small footprint portable laser profiler then flew over these inventory plots to develop a generic airborne LiDAR-based biomass equation (R² = 0.65, n = 207). The same airborne LiDAR system flew along four portions of orbits of the ICESat Geoscience Laser Altimeter System (GLAS). A square-root transformed equation was developed to predict airborne profiling LiDAR estimates of aboveground dry biomass from GLAS waveform parameters combined with an SRTM slope index (R² = 0.59, n = 1325).Using the 104,044 quality-filtered GLAS pulses obtained during autumn 2003 from 97 orbits over the study area, we then predicted aboveground dry biomass for the main vegetation areas of Québec as well as for the entire Province south of the treeline. Including cover type covariances both within and between GLAS orbits increased standard errors of the estimates by two to five times at the vegetation zone level and as much as threefold at the provincial level. Aboveground biomass for the whole study area averaged 39.0 ± 2.2 (standard error) Mg ha^? 1 and totalled 4.9 ± 0.3 Pg. Biomass distributions were 12.6% northern hardwoods, 12.6% northern mixedwood, 38.4% commercial boreal, 13% non-commercial boreal, 14.2% taiga, and 9.2% treed tundra. Non-commercial forests represented 36% of the estimated aboveground biomass, thus highlighting the importance of remote northern forests to C sequestration. This study has shown that space-based forest inventories of northern forests could be an efficient way of estimating the amount, distribution, and uncertainty of aboveground biomass and carbon stocks at large spatial scales.     

Using MODIS images to examine the surface extents and variations derived from the DEM and laser altimeter data in the Danjiangkou Reservoir,China

We used publicly available digital spatial datasets to study the area extents and their horizontal variations of two water bodies within the Danjiangkou Reservoir, China. Between 2003 and 2005, the water levels varied from 140 to 149 m above mean sea level as measured by the Geoscience Laser Altimeter System (GLAS). Detailed procedures to derive the horizontal extents and variations from the Moderate Resolution Imaging Spectroradiometer (MODIS) and Shuttle Radar Topography Mission (SRTM) digital elevation model (DEM) coupled with GLAS data and to verify the extents and variations spatially were provided. For the water bodies on the north and west, the surface water extents derived from four MODIS images varied between 174 and 218 km² and from 96 to 135 km², respectively. The extents by inundating the DEM using the GLAS data were 178–212 km² for the water body on the north and 104–118 km² for the water body on the west. The spatial verifications of surface water extents derived from the MODIS images versus DEM coupled with GLAS data agreed 83–93%. Within the ring areas between water/land boundaries at elevations of 140 and 147 m, and 140 and 149 m, the spatial agreement was 52–75%.     

Forest vertical structure from GLAS: An evaluation using LVIS and SRTM data

The Geoscience Laser Altimeter System (GLAS) on the Ice, Cloud and land Elevation Satellite (ICESat) is the first spaceborne lidar instrument for continuous global observation of the Earth. GLAS records a vertical profile of the returned laser energy from its footprint. To help understand the application of the data for forest spatial structure studies in our regional projects, an evaluation of the GLAS data was conducted using NASA's Laser Vegetation Imaging Sensor (LVIS) data in an area near NASA's Goddard Space Flight Center in Greenbelt, Maryland, USA. The tree height indices from airborne large-footprint lidars such as LVIS have been successfully used for estimation of forest structural parameters in many previous studies and served as truth in this study.The location accuracy of the GLAS footprints was evaluated by matching the elevation profile from GLAS with the Shuttle Radar Topography Mission (SRTM) DEM. The results confirmed the location accuracy of the GLAS geolocation, and showed a high correlation between the height of the scattering phase center from SRTM and the top tree height from GLAS data. The comparisons between LVIS and GLAS data showed that the GLAS waveform is similar to the aggregation of the LVIS waveforms within the GLAS footprint, and the tree height indices derived from the GLAS and LVIS waveforms were highly correlated. The best correlations were found between the 75% waveform energy quartiles of LVIS and GLAS (r² = 0.82 for October 2003 GLAS data, and r² = 0.65 for June 2005 GLAS data). The correlations between the 50% waveform energy quartiles of LVIS and GLAS were also high (0.77 and 0.66 respectively). The comparisons of the top tree height and total length of waveform of the GLAS data acquired in fall of 2003 and early summer of 2005 showed a several meter bias. Because the GLAS footprints from these two orbits did not exactly overlap, several other factors may have caused this observed difference, including difference of forest structures, seasonal difference of canopy structures and errors in identifying the ground peak of waveforms.     

TanDEM-X elevation model data for canopy height and aboveground biomass retrieval in a tropical peat swamp forest

It was demonstrated in the past that radar data is useful to estimate aboveground biomass due to their interferometric capability. Therefore, the potential of a globally available TanDEM-X digital elevation model (DEM) was investigated for aboveground biomass estimation via canopy height models (CHMs) in a tropical peat swamp forest. However, CHMs based on X-band interferometers usually require external terrain models. High accurate terrain models are not available on global scale. Therefore, an approach exclusively based on TanDEM-X and the decrease of accuracy compared to an approach utilizing a high accurate terrain model is assessed. In addition, the potential of X-band interferometric heights in tropical forests needs to be evaluated. Therefore, two CHMs were derived from an intermediate TanDEM-X DEM (iDEM; as a precursor for WorldDEM^TM) alone and in combination with lidar measurements used as terrain model. The analysis showed high accuracies (root mean square error [RMSE] = 5 m) for CHMs based on iDEM and reliable estimation of aboveground biomass. The iDEM CHM, exclusively based on TanDEM-X, achieved a poor R² of 0.2, nonetheless resulted in a cross-validated RMSE of 54 t ha^?1 (16%). The low R² suggested that the X-band height alone was not sufficient to estimate an accurate CHM, and thus the need for external terrain models was confirmed. A CHM retrieved from the difference of iDEM and an accurate lidar terrain model achieved a considerably higher correlation with aboveground biomass (R² = 0.68) and low cross-validated RMSE of 24.5 t ha^?1 (7.5%). This was higher or comparable to other aboveground biomass estimations in tropical peat swamp forests. The potential of X-band interferometric heights for CHM and biomass estimation was thus confirmed in tropical forest in addition to existing knowledge in boreal forests.     

Aboveground forest biomass derived using multiple dates of WorldView-2 stereo-imagery: quantifying the improvement in estimation accuracy

ABSTRACT

The aim of this study was to investigate the capabilities of two date satellite-derived image-based point clouds (IPCs) to estimate forest aboveground biomass (AGB). The data sets used include panchromatic WorldView-2 stereo-imagery with 0.46 m spatial resolution representing 2014 and 2016 and a detailed digital elevation model derived from airborne laser scanning data. Altogether, 332 field sample plots with an area of 256 m² were used for model development and validation. Predictors describing forest height, density, and variation in height were extracted from the IPC 2014 and 2016 and used in k-nearest neighbour imputation models developed with sample plot data for predicting AGB. AGB predictions for 2014 (AGB₂₀₁₄) were projected to 2016 using growth models (AGB_{Projected_2016}) and combined with the AGB estimates derived from the 2016 data (AGB₂₀₁₆). AGB prediction model developed with 2014 data was also applied to 2016 data (AGB_{2016_pred2014}). Based on our results, the change in the 90^th percentile of height derived from the WorldView-2 IPC was able to characterize forest height growth between 2014 and 2016 with an average growth of 0.9 m. Features describing canopy cover and variation in height derived from the IPC were not as consistent. The AGB₂₀₁₆ had a bias of ?7.5% (?10.6 Mg ha^?1) and root mean square error (RMSE) of 26.0% (36.7 Mg ha^?1) as the respective values for AGB_{Projected_2016} were 7.0% (9.9 Mg ha^?1) and 21.5% (30.8 Mg ha^?1). AGB_{2016_pred2014} had a bias of ?19.6% (?27.7 Mg ha^?1) and RMSE of 33.2% (46.9 Mg ha^?1). By combining predictions of AGB₂₀₁₆ and AGB_{Projected_2016} at sample plot level as a weighted average, we were able to decrease the bias notably compared to estimates made on any single date. The lowest bias of ?0.25% (?0.4 Mg ha^?1) was obtained when equal weights of 0.5 were given to the AGB_{Projected_2016} and AGB₂₀₁₆ estimates. Respectively, RMSE of 20.9% (29.5 Mg ha^?1) was obtained using equal weights. Thus, we conclude that combination of two date WorldView-2 stereo-imagery improved the reliability of AGB estimates on sample plots where forest growth was the only change between the two dates.     

Estimating plot volume using lidar height and intensity distributional parameters

This study explored the feasibility of height distributional metrics and intensity values extracted from low-density airborne light detection and ranging (lidar) data to estimate plot volumes in dense Korean pine (Pinus koraiensis) plots. Multiple linear regression analyses were performed using lidar height and intensity distributional metrics. The candidate variables for predicting plot volume were evaluated using three data sets: total, canopy, and integrated lidar height and intensity metrics. All intensities of lidar returns used were corrected by the reference distance. Regression models were developed using each data set, and the first criterion used to select the best models was the corrected Akaike Information Criterion (AIC_c). The use of three data sets was statistically significant at R² = 0.75 (RMSE = 52.17 m³ ha^?1), R² = 0.84 (RMSE = 45.24 m³ ha^?1), and R² = 0.91 (RMSE = 31.48 m³ ha^?1) for total, canopy, and integrated lidar distributional metrics, respectively. Among the three data sets, the integrated lidar metrics-derived model showed the best performance for estimating plot volumes, improving errors up to 42% when compared to the other two data sets. This is attributed to supplementing variables weighted and biased to upper limits in dense plots with more statistical variables that explain the lower limits. In all data sets, intensity metrics such as skewness, kurtosis, standard deviation, minimum, and standard error were employed as explanatory variables. The use of intensity variables improved the accuracy of volume estimation in dense forests compared to prior research. Correction of the intensity values contributed up to a maximum of 58% improvement in volume estimation when compared to the use of uncorrected intensity values (R² = 0.78, R² = 0.53, and R² = 0.63 for total, canopy, and integrated lidar distributional metrics, respectively). It is clear that the correction of intensity values is an essential step for the estimation of forest volume.     

Characterizing forest canopy structure with lidar composite metrics and machine learning

A lack of reliable observations for canopy science research is being partly overcome by the gradual use of lidar remote sensing. This study aims to improve lidar-based canopy characterization with airborne laser scanners through the combined use of lidar composite metrics and machine learning models. Our so-called composite metrics comprise a relatively large number of lidar predictors that tend to retain as much information as possible when reducing raw lidar point clouds into a format suitable as inputs to predictive models of canopy structural variables. The information-rich property of such composite metrics is further complemented by machine learning, which offers an array of supervised learning models capable of relating canopy characteristics to high-dimensional lidar metrics via complex, potentially nonlinear functional relationships. Using coincident lidar and field data over an Eastern Texas forest in USA, we conducted a case study to demonstrate the ubiquitous power of the lidar composite metrics in predicting multiple forest attributes and also illustrated the use of two kernel machines, namely, support vector machine and Gaussian processes (GP). Results show that the two machine learning models in conjunction with the lidar composite metrics outperformed traditional approaches such as the maximum likelihood classifier and linear regression models. For example, the five-fold cross validation for GP regression models (vs. linear/log-linear models) yielded a root mean squared error of 1.06 (2.36) m for Lorey's height, 0.95 (3.43) m for dominant height, 5.34 (8.51) m²/ha for basal area, 21.4 (40.5) Mg/ha for aboveground biomass, 6.54 (9.88) Mg/ha for belowground biomass, 0.75 (2.76) m for canopy base height, 2.2 (2.76) m for canopy ceiling height, 0.015 (0.02) kg/m³ for canopy bulk density, 0.068 (0.133) kg/m² for available canopy fuel, and 0.33 (0.39) m²/m² for leaf area index. Moreover, uncertainty estimates from the GP regression were more indicative of the true errors in the predicted canopy variables than those from their linear counterparts. With the ever-increasing accessibility of multisource remote sensing data, we envision a concomitant expansion in the use of advanced statistical methods, such as machine learning, to explore the potentially complex relationships between canopy characteristics and remotely-sensed predictors, accompanied by a desideratum for improved error analysis.     

Physically based vertical vegetation structure retrieval from ICESat data: Validation using LVIS in White Mountain National Forest, New Hampshire, USA

Vegetation structure retrieval accuracies from spaceborne Geoscience Laser Altimeter System (GLAS) on the Ice, Cloud and land Elevation Satellite (ICESat) data are affected by surface topography, background noise and sensor saturation. This study uses a physical approach to remove surface topography effect from lidar returns to retrieve vegetation height from ICESat/GLAS data over slope terrains. Slope-corrected vegetation heights from ICESat/GLAS data were compared to airborne Laser Vegetation Imaging Sensor (LVIS) (20 m footprint size) and small-footprint lidar data collected in White Mountain National Forest, NH. Impact of slope on LVIS vegetation height estimates was assessed by comparing LVIS height before and after slope correction with small-footprint discrete-return lidar and field data.Slope-corrected GLAS vegetation heights match well with 98 percentile heights from small-footprint lidar (R² = 0.77, RMSE = 2.2 m) and top three LVIS mean (slope-corrected) heights (R² = 0.64, RMSE = 3.7 m). Impact of slope on LVIS heights is small, however, comparison of LVIS heights (without slope correction) with either small footprint lidar or field data indicates that our scheme improves the overall LVIS height accuracy by 0.4-0.7 m in this region. Vegetation height can be overestimated by 3 m over a 15° slope without slope correction. More importantly, both slope-corrected GLAS and LVIS height differences are independent of slope. Our results demonstrate the effectiveness of the physical approach to remove surface topography from large footprint lidar data to improve accuracy of maximum vegetation height estimates.GLAS waveforms were compared to aggregated LVIS waveforms in Bartlett Experimental Forest, NH, to evaluate the impact of background noise and sensor saturation on vegetation structure retrievals from ICESat/GLAS. We found that GLAS waveforms with sensor saturation and low background noise match well with aggregated LVIS waveforms, indicating these waveforms capture vertical vegetation structure well. However, waveforms with large noise often lead to mismatched waveforms with LVIS and underestimation of waveform extent and vegetation height. These results demonstrate the quality of ICESat/GLAS vegetation structure estimates.     

Modeling the height of young forests regenerating from recent disturbances in Mississippi using Landsat and ICESat data

Many forestry and earth science applications require spatially detailed forest height data sets. Among the various remote sensing technologies, lidar offers the most potential for obtaining reliable height measurement. However, existing and planned spaceborne lidar systems do not have the capability to produce spatially contiguous, fine resolution forest height maps over large areas. This paper describes a Landsat-lidar fusion approach for modeling the height of young forests by integrating historical Landsat observations with lidar data acquired by the Geoscience Laser Altimeter System (GLAS) instrument onboard the Ice, Cloud, and land Elevation (ICESat) satellite. In this approach, “young” forests refer to forests reestablished following recent disturbances mapped using Landsat time-series stacks (LTSS) and a vegetation change tracker (VCT) algorithm. The GLAS lidar data is used to retrieve forest height at sample locations represented by the footprints of the lidar data. These samples are used to establish relationships between lidar-based forest height measurements and LTSS-VCT disturbance products. The height of “young” forest is then mapped based on the derived relationships and the LTSS-VCT disturbance products. This approach was developed and tested over the state of Mississippi. Of the various models evaluated, a regression tree model predicting forest height from age since disturbance and three cumulative indices produced by the LTSS-VCT method yielded the lowest cross validation error. The R² and root mean square difference (RMSD) between predicted and GLAS-based height measurements were 0.91 and 1.97 m, respectively. Predictions of this model had much higher errors than indicated by cross validation analysis when evaluated using field plot data collected through the Forest Inventory and Analysis Program of USDA Forest Service. Much of these errors were due to a lack of separation between stand clearing and non-stand clearing disturbances in current LTSS-VCT products and difficulty in deriving reliable forest height measurements using GLAS samples when terrain relief was present within their footprints. In addition, a systematic underestimation of about 5 m by the developed model was also observed, half of which could be explained by forest growth that occurred between field measurement year and model target year. The remaining difference suggests that tree height measurements derived using waveform lidar data could be significantly underestimated, especially for young pine forests. Options for improving the height modeling approach developed in this study were discussed.     

Estimating standing biomass in papyrus (Cyperus papyrus L.) swamp: exploratory of in situ hyperspectral indices and random forest regression

Accurate estimates of papyrus (Cyperus papyrus) biomass are critical for an efficient papyrus swamp monitoring and management system. The objective of this study was to test the utility of random forest (RF) regression and two narrow-band vegetation indices in estimating above-ground biomass (AGB) for complex and densely vegetated swamp canopies. The normalized difference vegetation index (NDVI) and enhanced vegetation index (EVI) were calculated from field spectrometry data and fresh AGB was measured in 82 quadrats at three different areas in the iSimangaliso Wetland Park, South Africa. NDVI was calculated from all possible band combinations of the electromagnetic spectrum (350 and 2500 nm), while EVI was calculated from possible band combinations in the blue, red, and near infrared of the spectrum. Backward feature elimination and RF regression were used as variable selection and modelling techniques to predict papyrus AGB. Results showed that the effective portions of electromagnetic spectrum for estimation AGB of papyrus swamp were located within the blue, red, red-edge, and near-infrared regions. The three best selected EVIs were computed from bands located at (i) 445, 682, and 829 nm, (ii) 497, 676, and 1091 nm, and (iii) 495, 678, and 1120 nm. These indices produced better predictive accuracies (R² = 0.90; root mean square error of prediction (RMSEP) = 0.289 kg m^?2; 7.99% of the mean) than the best selected NDVIs (R² = 0.85; RMSEP = 0.343 kgm^?2; 9.49% of the mean) that were calculated from bands located at (i) 739 and 829 nm, (ii) 739 and 814 nm, (iii) 744 and 789 nm, and (iv) 734 and 909 nm. The results of the present study demonstrate the utility of narrow-band vegetation indices (EVI and NDVI) and RF regression in estimating papyrus AGB at high density, a previously challenging task with broadband satellite sensors.     

Comparative analysis of SRTM–NED vegetation canopy height to LIDAR‐derived vegetation canopy metrics

Vegetation canopy heights derived from the SRTM 30 m grid DEM minus USGS National Elevation Data (NED) DTM were compared to three vegetation metrics derived from a medium footprint LIDAR data (LVIS) for the US Sierra Nevada forest in California. Generally the SRTM minus NED was found to underestimate the vegetation canopy height. Comparing the SRTM–NED‐derived heights as a function of the canopy percentile height (shape/vertical structure) derived from LVIS, the SRTM SAR signal was found to penetrate, on average, into about 44% of the canopy and 85% after adjustment of the data. On the canopy type analysis, it was found that the SRTM phase scattering centres occurred at 60% for red fir, 53% for Sierra mixed conifer, 50% for ponderosa pine and 50% for montane hardwood‐conifer. Whereas analysing the residual errors of the SRTM–NED minus the LVIS‐derived canopy height as a function of LVIS canopy height and cover it was observed that the residuals generally increase with increasing canopy height and cover. Likewise, the behaviour of the RMSE as a function of canopy height and cover was observed to initially increase with canopy height and cover but saturates at 50 m canopy height and 60% canopy cover. On the other hand, the behaviour of the correlation coefficient as a function of canopy height and cover was found to be high at lower canopy height (<15 m) and cover (<20%) and decrease rapidly making a depression at medium canopy heights (>15 m and <50 m) and cover (>20% and <50%). It then increases with increasing canopy height and cover yielding a plateau at canopies higher than 50 m and cover above 70%.     

Estimation of tropical rain forest aboveground biomass with small-footprint lidar and hyperspectral sensors

Tropical forests are an important component of the global carbon balance, yet there is considerable uncertainty in estimates of their carbon stocks and fluxes, which are typically estimated through analysis of aboveground biomass in field plots. Remote sensing technology is critical for assessing fine-scale spatial variability of tropical forest biomass over broad spatial extents. The goal of our study was to evaluate relatively new technology, small-footprint, discrete-return lidar and hyperspectral sensors, for the estimation of aboveground biomass in a Costa Rican tropical rain forest landscape. We derived a suite of predictive metrics for field plots: lidar metrics were calculated from plot vertical height profiles and hyperspectral metrics included fraction of spectral mixing endmembers and narrowband indices that respond to photosynthetic vegetation, structure, senescence, health and water and lignin content. We used single- and two-variable linear regression analyses to relate lidar and hyperspectral metrics to aboveground biomass of plantation, managed parkland and old-growth forest plots. The best model using all 83 biomass plots included two lidar metrics, plot-level mean height and maximum height, with an r² of 0.90 and root-mean-square error (RMSE) of 38.3 Mg/ha. When the analysis was constrained to plantation plots, which had the most accurate field data, the r² of the model increased to 0.96, with RMSE of 10.8 Mg/ha (n = 32). Hyperspectral metrics provided lower accuracy in estimating biomass than lidar metrics, and models with a single lidar and hyperspectral metric were no better than the best model using two lidar metrics. These results should be viewed as an initial assessment of using these combined sensors to estimate tropical forest biomass; hyperspectral data were reduced to nine indices and three spectral mixture fractions, lidar data were limited to first-return canopy height, sensors were flown only once at different seasons, and we explored only linear regression for modeling. However, this study does support conclusions from studies at this and other climate zones that lidar is a premier instrument for mapping biomass (i.e., carbon stocks) across broad spatial scales.     

Biophysical estimation in tropical forests using JERS‐1 SAR and VNIR imagery. II. Aboveground woody biomass

Accurate estimates of aboveground biomass in tropical forests are important in carbon sequestration and global change studies. Tropical forest biomass estimation with microwave remote sensing is limited because of the strong scattering and attenuation properties of the green canopy. In this study a microwave/optical synergistic model was developed to quantify these effects to Synthetic Aperture Radar (SAR) signals and to better estimate woody structures, which are closely related to aboveground biomass. With a Leaf Area Index (LAI) retrieved from Japan Earth Resources Satellite (JERS)‐1 Very Near Infrared Radiometer (VNIR) imagery, leaf scattering and attenuation to woody scattering were quantified and removed from the total backscatter in a modified canopy scattering model. Woody scattering showed high sensitivity to biomass >100 tonnes/ha in tropical forests. Tree height and stand density were derived from the JERS‐1 SAR image with a root mean square error (RMSE) of 4 m and 161 trees/ha, respectively. Aboveground biomass was calculated using a general allometric equation. Biomass in secondary dry dipterocarps (Dipterocarpaceae family of tropical lowland deciduous trees) was overestimated. The modelled biomass in mixed deciduous and dry evergreen forests fit better with ground measurements. In mountainous areas with steep slopes, the topographic effects in the SAR image could not be properly corrected and therefore the results are unreliable.     

Allometric scaling theory-based maximum forest tree height and biomass estimation in the Three Gorges reservoir region using multi-source remote-sensing data

Most terrestrial carbon is stored in forest biomass, which plays an important role in local, regional, and global climate change. Monitoring of forests and their status, and accurate estimation of forest biomass are important in mitigating the impacts of climate change. Empirical models developed using remote-sensing and field-measured forest data are commonly used to estimate forest biomass. In the present study, we used a mechanistic model to estimate height and biomass in the Three Gorges reservoir region (China) based on the allometric scale and resource limits (ASRL) model. The forests in the Three Gorges reservoir region are important and unique in view of the vertical distribution of vegetation and mixed needleleaf. Detailed information about the forest in this region is available from the Geoscience Laser Altimeter System (GLAS) and field measurements from 714 forest plots. The ASRL model parameters were adjusted using GLAS-derived forest tree height to reduce the deviation between modelled and observed forest height. The predicted maximum forest tree height from the optimized ASRL model was compared to measured tree heights, and a good correlation (R² = 0.566) was found. The allometric scale function between forest height and diameter at breast height (DBH) is developed and the maximum forest tree height from the optimized ASRL model transferred to DBH. Moreover, the forest biomass was estimated from DBH according to the allometric scale function that was determined using DBH and biomass data. The results of maximum forest biomass using the ASRL model and the allometric scale function show a good accuracy (R² = 0.887) in the Three Gorges reservoir region. Here, we present the forest biomass estimation approach following allometric theory for accurate estimation of maximum forest tree height and biomass. The proposed approach can be applied to forest species in all types of environmental conditions.     

Estimating aboveground biomass using Landsat 7 ETM+ data across a managed landscape in northern Wisconsin, USA

Aboveground biomass (AGB; Mg/ha) is defined in this study as a biomass of growing stock trees greater than 2.5 cm in diameter at breast height (dbh) for stands >5 years and all trees taller than 1.3 m for stands <5 years. Although AGB is an important variable for evaluating ecosystem function and structure across the landscape, such estimates are difficult to generate without high-resolution satellite data. This study bridges the application of remote sensing techniques with various forest management practices in Chequamegon National Forest (CNF), Wisconsin, USA by producing a high-resolution stand age map and a spatially explicit AGB map. We coupled AGB values, calculated from field measurements of tree dbh, with various vegetation indices derived from Landsat 7 ETM+ data through multiple regression analyses to produce an initial biomass map. The initial biomass map was overlaid with a land-cover map to generate a stand age map. Biomass threshold values for each age category (e.g., young, intermediate, and mature) were determined through field observations and frequency analysis of initial biomass estimates by major cover types. We found that AGB estimates for hardwood forests were strongly related to stand age and near-infrared reflectance (r²=0.95) while the AGB for pine forests was strongly related to the corrected normalized difference vegetation index (NDVIc; r²=0.86). Separating hardwoods from pine forests improved the AGB estimates in the area substantially, compared to overall regression (r²=0.82). Our AGB results are comparable to previously reported values in the area. The total amount of AGB in the study area for 2001 was estimated as 3.3 million metric tons (dry weight), 76.5% of which was in hardwood and mixed hardwood/pine forests. AGB ranged from 1 to 358 Mg/ha with an average of 70 and a standard deviation of 54 Mg/ha. The AGB class with the highest percentage (16.1%) was between 81 and 100 Mg/ha. Forests with biomass values >200 Mg/ha accounted for less than 3% of the study area and were usually associated with mature hardwood forests. Estimated AGB was validated using independent field measurements (R²=0.67, p<0.001). The AGB and age maps can be used as baseline information for future landscape level studies such as quantifying the regional carbon budget, accumulating fuel, or monitoring management practices.     

Forest inventory stand height estimates from very high spatial resolution satellite imagery calibrated with lidar plots

Many areas of forest across northern Canada are challenging to monitor on a regular basis as a result of their large extent and remoteness. Although no forest inventory data typically exist for these northern areas, detailed and timely forest information for these areas is required to support national and international reporting obligations. We developed and tested a sample-based approach that could be used to estimate forest stand height in these remote forests using panchromatic Very High Spatial Resolution (VHSR, < 1 m) optical imagery and light detection and ranging (lidar) data. Using a study area in central British Columbia, Canada, to test our approach, we compared four different methods for estimating stand height using stand-level and crown-level metrics generated from the VHSR imagery. ‘Lidar plots’ (voxel-based samples of lidar data) are used for calibration and validation of the VHSR-based stand height estimates, similar to the way that field plots are used to calibrate photogrammetric estimates of stand height in a conventional forest inventory or to make empirical attribute estimates from multispectral digital remotely sensed data. A k-nearest neighbours (k-NN) method provided the best estimate of mean stand height (R ² = 0.69; RMSE = 2.3 m, RMSE normalized by the mean value of the estimates (RMSE-%) = 21) compared with linear regression, random forests, and regression tree methods. The approach presented herein demonstrates the potential of VHSR panchromatic imagery and lidar to provide robust and representative estimates of stand height in remote forest areas where conventional forest inventory approaches are either too costly or are not logistically feasible. While further evaluation of the methods is required to generalize these results over Canada to provide robust and representative estimation, VHSR and lidar data provide an opportunity for monitoring in areas for which there is no detailed forest inventory information available.     

Retrieving vegetation height of forests and woodlands over mountainous areas in the Pacific Coast region using satellite laser altimetry

The challenge to retrieve canopy height from large-footprint satellite lidar waveforms over mountainous areas is formidable given the complex interaction of terrain and vegetation. This study explores the potential of GLAS (Geoscience Laser Altimeter System) for retrieving maximum canopy height over mountainous areas in the Pacific Coast region, including two conifers sites of tall and closed canopy and one broadleaf woodland site of shorter and sparse canopy. Both direct methods and statistical models are developed and tested using spatially extensive coincident airborne lidar data. The major findings include: 1) the direct methods tend to overestimate the canopy height and are complicated by the identification of waveform signal start and terrain ground elevation, 2) the exploratory data analysis indicates that the edge-extent linear regression models have better generalizability than the edge-extent nonlinear models at the inter-site level, 3) the inter-site level test with mixed-effects models reveals that the edge-extent linear models have statistically-justified generalizability between the two conifer sites but not between the conifer and woodland sites, 4) the intra-site level test indicates that the edge-extent linear models have statistically-justified generalizability across different vegetation community types within any given site; this, combined with 3), unveils that the statistical modeling of maximum canopy height over large areas with edge-extent linear models only need to consider broad vegetation differences (such as woodlands versus conifer forests instead of different vegetation communities within woodlands or conifer forests), and 5) the simulations indicate that the errors and uncertainty in canopy height estimation can be significantly reduced by decreasing the footprint size. It is recommended that the footprint size of the next-generation satellite lidar systems be at least 10 m or so if we want to achieve meter-level accuracy of maximum canopy height estimation using direct and statistical methods.