Faculty Publications - Daniels Faculty of Architecture, Landscape, and Design

Permanent URI for this collectionhttps://hdl.handle.net/1807/108075

Browse

Now showing 1 - 9 of 9

A mathematical framework to describe the effect of beam incidence angle on metrics derived from airborne LiDAR: The case of forest canopies approaching turbid medium behaviour
(Elsevier, 2018-05) Roussel, Jean-Romain; Béland, Martin; Caspersen, John; Achim, Alexis
Airborne laser scanning (LiDAR) is used in forest inventories to quantify stand structure with three-dimensional point clouds. However, the 3D distribution of the point clouds depends not only on stand structure, but also on scan angle, because the probability for an oblique beam to be reflected by the canopy increases with the distance it must travel through the canopy. Thus, the canopy appears to increase in density as the incidence angle increases, all else being equal. The resulting variation between and within datasets can induce bias in LiDAR metrics derived from the vertical distribution of points. In this study, we modelled the effect of scan angle on the vertical structure of the point clouds to predict the bias of metrics derived from points sampled off-nadir. Comparison with paired observations from different flightlines (off- and at-nadir observations of the same point) demonstrated that the model accurately reproduced the bias of metrics calculated for a northern hardwood forest with relatively continuous canopy. Thus, the model could be used to correct the bias of LiDAR metrics, and provides a mathematical framework that could be used to inform the selection of maximum incidence angle in LiDAR surveys, considering the trade-off between decreasing acquisition costs and obtaining unbiased measurements.
Removing bias from LiDAR-based estimates of canopy height: Accounting for the effects of pulse density and footprint size
(Elsevier, 2017-09-01) Roussel, Jean-Romain; Caspersen, John; Béland, Martin; Thomas, Sean; Achim, Alexis
Airborne laser scanning (LiDAR) is used in forest inventories to quantify stand structure with three dimensional point clouds. However, the structure of point clouds depends not only on stand structure, but also on the LiDAR instrument, its settings, and the pattern of flight. The resulting variation between and within datasets (particularly variation in pulse density and footprint size) can induce spurious variation in LiDAR metrics such as maximum height (hmax) and mean height of the canopy surface model (Cmean). In this study, we first compare two LiDAR datasets acquired with different parameters, and observe that hmax and Cmean are 56 cm and 1.0 m higher, respectively, when calculated using the high-density dataset with a small footprint. Then, we present a model that explains the observed bias using probability theory, and allows us to recompute the metrics as if the density of pulses were infinite and the size of the two footprints were equivalent. The model is our first step in developing methods for correcting various LiDAR metrics that are used for area-based prediction of stand structure. Such methods may be particularly useful for monitoring forest growth over time, given that acquisition parameters often change between inventories.
Reproductive costs in Acer saccharum: exploring size-dependent relations between seed production and branch extension
(Springer, 2017-08) Hossain, Shaik Md. Yousuf; Caspersen, John P.; Thomas, Sean C.
Life-history theory predicts that reproductive allocation should increase with age and size once plants reach reproductive maturity. This suggests that there may also be a subsequent decline in somatic growth as plants become larger or older. However, few studies have examined how the relationship between branch extension growth and reproduction varies with size or age in the longest-lived plants: trees. Using a mobile lift for canopy access, we retrospectively measured branch extension growth before, during and after two (between 2011 and 2013) Acer saccharum mast events (the synchronous production of many seeds at long intervals), quantifying seed production per internode and internode length. Branch extension was reduced by 24 and 36%, respectively, in 2011 and 2013 relative to non-mast years, consistent with the expectation that increased reproductive allocation comes at the cost of allocation to growth. Internode length decreased from 8 to 3 cm year−1 as seed production increased from zero to 17 seeds year−1; a similar decrease was observed at the whole-tree level using average internode extension rates and seed production per tree. Seed production alone was the most parsimonious predictor of branch extension growth, with no independent effect of tree size, suggesting that it is the increase in reproductive allocation, rather than an increase in tree size per se, that drives the decline in branch extension rates. The slope of the relationship between branch extension and reproduction did not vary with tree size, suggesting that there was no increase in the somatic cost of reproduction with tree size. We also found no evidence for lag effects of reproduction on extension growth in subsequent years. Overall, these results suggest that reproductive allocation assessed at the shoot level increases with tree size and is a major driver of the ontogenetic decline in branch extension growth.
Fully constrained linear spectral unmixing based global shadow compensation for high resolution satellite imagery of urban areas
(Elsevier, 2015-06) Yang, Jian; He, Yuhong; Caspersen, John
Shadows commonly exist in high resolution satellite imagery, particularly in urban areas, which is a combined effect of low sun elevation, off-nadir viewing angle, and high-rise buildings. The presence of shadows can negatively affect image processing, including land cover classification, mapping, and object recognition due to the reduction or even total loss of spectral information in shadows. The compensation of spectral information in shadows is thus one of the most important preprocessing steps for the interpretation and exploitation of high resolution satellite imagery in urban areas. In this study, we propose a new approach for global shadow compensation through the utilization of fully constrained linear spectral unmixing. The basic assumption of the proposed method is that the construction of the spectral scatter plot in shadows is analogues to that in non-shadow areas within a two-dimension spectral mixing space. In order to ensure the continuity of land covers, a smooth operator is further used to refine the restored shadow pixels on the edge of non-shadow and shadow areas. The proposed method is validated using the WorldView-2 multispectral imagery collected from downtown Toronto, Ontario, Canada. In comparison with the existing linear-correlation correction method, the proposed method produced the compensated shadows with higher quality.
Estimation of forest structural and compositional variables using ALS data and multi-seasonal satellite imagery
(Elsevier, 2019-06) Shang, Chen; Treitz, Paul; Caspersen, John; Jones, Trevor
Advanced forest resource inventory (FRI) information is of critical importance for sustainable forest management. FRIs are dependent on remote sensing data and processing methods, along with field calibration/validation to generate cost-effective options for modelling forest inventory and biophysical variables over large areas. The objective of this study was to examine the impact of combining multi-seasonal multispectral satellite imagery with airborne laser scanning (ALS) data for estimating basal area, species mixture and stem density for an uneven-aged tolerant hardwood forest in Ontario, Canada. Using random forest (RF) regression as a non-parametric diagnostic technique, three multispectral optical sensors (i.e., Landsat-5 TM, Sentinel-2 A and WorldView-2) were compared to examine the most cost-effective sensor configuration for modelling FRI variables. The contribution of spectral predictors derived from these optical sensors as well as ALS height and intensity metrics were evaluated using RF variable importance. As part of our variable selection framework, all predictor variables were grouped into relatively independent clusters using a hierarchical variable clustering technique, which revealed the distinctiveness between information contained in spectral predictors, height- and intensity-based metrics. This indicates that ALS intensity data carry unique information complementary to passive near-infrared data for forest characterization. ALS data alone did not result in accurate models for basal area and species mixture, but predictive accuracies were improved significantly with the addition of spectral predictors. Compared to single-date images, multi-seasonal imagery proved to be more accurate for modelling FRI variables, especially when combined with ALS data. Despite its limited spatial resolution, Sentinel-2 A was found to be the most cost-effective image source for enhancing ALS-based FRI models. Using variables identified by the variable selection procedure, best subsets regression outperformed the RF models developed for diagnostic analysis, resulting in a suite of accurate and parsimonious predictive models, with coefficients of determination of 0.73, 0.90 and 0.67, for basal area, species mixture, and stem density, respectively.
A discrepancy measure for segmentation evaluation from the perspective of object recognition
(Elsevier, 2015-03) Yang, Jian; He, Yuhong; Caspersen, John; Jones, Trevor
Within the framework of geographic object-based image analysis (GEOBIA), segmentation evaluation is one of the most important components and thus plays a critical role in controlling the quality of GEOBIA workflow. Among a variety of segmentation evaluation methods and criteria, discrepancy measurement is believed to be the most useful and is therefore one of the most commonly employed techniques in many applications. Existing measures have largely ignored the importance of object recognition in segmentation evaluation. In this study, a new discrepancy measure of segmentation evaluation index (SEI) redefines the corresponding segment using a two-sided 50% overlap instead of one-sided 50% overlap that has been commonly used. The effectiveness of SEI is further investigated using the schematic segmentation cases and remote sensing images. Results demonstrate that the proposed SEI outperforms the other two existing discrepancy measures, Euclidean Distance 2 (ED2) and Euclidean Distance 3 (ED3), both in terms of object recognition accuracy and identification of detailed segmentation differences.
Comparing the life cycle costs of using harvest residue as feedstock for small- and large-scale bioenergy systems (part II)
(Elsevier, 2015-06) Cleary, Julian; Wolf, Derek P.; Caspersen, John P.
In part II of our two-part study, we estimate the nominal electricity generation and GHG (greenhouse gas) mitigation costs of using harvest residue from a hardwood forest in Ontario, Canada to fuel (1) a small-scale (250 kWe) combined heat and power wood chip gasification unit and (2) a large-scale (211 MWe) coal-fired generating station retrofitted to combust wood pellets. Under favorable operational and regulatory conditions, generation costs are similar: 14.1 and 14.9 cents per kWh (c/kWh) for the small- and large-scale facilities, respectively. However, GHG mitigation costs are considerably higher for the large-scale system: $159/tonne of CO2 eq., compared to $111 for the small-scale counterpart. Generation costs increase substantially under existing conditions, reaching: (1) 25.5 c/kWh for the small-scale system, due to a regulation mandating the continual presence of an operating engineer; and (2) 22.5 c/kWh for the large-scale system due to insufficient biomass supply, which reduces plant capacity factor from 34% to 8%. Limited inflation adjustment (50%) of feed-in tariff rates boosts these costs by 7% to 11%. Results indicate that policy generalizations based on scale require careful consideration of the range of operational/regulatory conditions in the jurisdiction of interest. Further, if GHG mitigation is prioritized, small-scale systems may be more cost-effective.
Assessing Coarse Woody Debris Nutrient Dynamics in Managed Northern Hardwood Forests Using a Matrix Transition Model
(Springer, 2020-04) Gorgolewski, Adam; Rudz, Philip; Jones, Trevor; Basiliko, Nathan; Caspersen, John
Coarse woody debris (CWD) is a dynamic source of nutrients in managed forests of eastern North America. The temporal patterns of nutrient export from CWD are challenging to study, and efficient methods are lacking. We made empirical measurements of CWD density, volume, and nutrient concentrations in 5 stages of decay, and paired them with a decay class transition model to project the long-term nutrient dynamics of CWD in a managed northern hardwood forest. The model was used to describe stand-level changes in CWD nutrient pools over 40 years following a selection harvest, and to compare CWD nutrient pools in managed and unmanaged stands. The C content of CWD decreased throughout decay, and mirrored density losses. N, P, and Ca content increased throughout decay, Mg content remained relatively constant, and K was rapidly lost. At the stand level, despite a rapid loss of mass and density, the model projected an initial gain in total N, P and Ca stored in CWD during the first 4–8 years after harvest, whereas net C, Mg, and K began to decrease immediately. The average volume, mass, C and K stocks of CWD in managed stands were approximately 10% lower than unmanaged stands, and N, P, Ca, and Mg were up to 16% lower. This is the first study to use a decay class transition model to study the dynamics of nutrients other than C, and the model serves as a template upon which other models of CWD decay can be built.
A simple area-based model for predicting airborne LiDAR first returns from stem diameter distributions: an example study in an uneven-aged, mixed temperate forest
(Canadian Science Publishing, 2015-10) Spriggs, Rebecca A.; Vanderwel, Mark C.; Jones, Trevor A.; Caspersen, John P.; Coomes, David A.
Tree size distributions are of fundamental importance in forestry. Airborne laser scanning (i.e., light detection and ranging, LiDAR) provides high-resolution information on canopy structure and may have potential as a tool for mapping and monitoring tree stem diameter distributions across forest landscapes. We present an area-based allometric model (with three levels of species specificity) that links ground-based plot data to the height distribution of LiDAR first returns, demonstrating the approach with survey data from a mixed, uneven-aged forest in central Ontario, Canada. Our model translates stem diameters into estimates of exposed crown area within 1 m height intervals; we then compared those estimates with the height distribution of LiDAR first returns. This basic approach gave reasonable goodness of fits (root mean squared error = 32%), but accuracy was improved by adding mechanistic features (root mean squared error = 17%) to adjust crown shapes and crown permeability and allow for crown overlap and gaps. The model showed no bias in predicting LiDAR returns in the mid to upper canopy (18–30 m) but tended to underestimate the returns from the understory level (2–8 m) and overestimate returns from the ground level and lower canopy (8–18 m). Our model represents an important contribution towards the remote mapping of tree size distributions by showing that LiDAR first returns can be accurately predicted from standard plot data via the inclusion of a few fundamental canopy properties.

Browse

Recent Submissions