L. Monika Moskal

Retrieving Leaf Area Index (LAI) Using Remote Sensing: Theories, Methods and Sensors

The ability to accurately and rapidly acquire leaf area index (LAI) is an indispensable component of process-based ecological research facilitating the understanding of gas-vegetation exchange phenomenon at an array of spatial scales from the leaf to the landscape. However, LAI is difficult to directly acquire for large spatial extents due to its time consuming and work intensive nature. Such efforts have been significantly improved by the emergence of optical and active remote sensing techniques. This paper reviews the definitions and theories of LAI measurement with respect to direct and indirect methods. Then, the methodologies for LAI retrieval with regard to the characteristics of a range of remotely sensed datasets are discussed. Remote sensing indirect methods are subdivided into two categories of passive and active remote sensing, which are further categorized as terrestrial, aerial and satellite-born platforms. Due to a wide variety in spatial resolution of remotely sensed data and the requirements of ecological modeling, the scaling issue of LAI is discussed and special consideration is given to extrapolation of measurement to landscape and regional levels.

Retrieving Forest Inventory Variables with Terrestrial Laser Scanning (TLS) in Urban Heterogeneous Forest

We present the point cloud slicing (PCS) algorithm, to post process point cloud data (PCD) from terrestrial laser scanning (TLS). We then test this tool for forest inventory application in urban heterogeneous forests. The methodology was based on a voxel data structure derived from TLS PCD. We retrieved biophysical tree parameters including diameter at breast height (DBH), tree height, basal area, and volume. Our results showed that TLS-based metrics explained 91.17% (RMSE = 9.1739 cm, p < 0.001) of the variation in DBH at individual tree level. Though the scanner generated a high-density PCD, only 57.27% (RMSE = 0.7543 m, p < 0.001) accuracy was achieved for predicting tree heights in these very heterogeneous stands. Furthermore, we developed a voxel-based TLS volume estimation method. Our results showed that PCD generated from TLS single location scans only captures 18% of the total tree volume due to an occlusion effect; yet there are significant relationships between the TLS data and field measured parameters for DBH and height, giving promise to the utility of a side scanning approach. Using our method, a terrestrial LiDAR-based inventory, also applicable to mobile- or vehicle-based laser scanning (MLS or VLS), was produced for future calibration of Aerial Laser Scanning (ALS) data and urban forest canopy assessments.

Tree Species Detection Accuracies Using Discrete Point Lidar and Airborne Waveform Lidar

Species information is a key component of any forest inventory. However, when performing forest inventory from aerial scanning Lidar data, species classification can be very difficult. We investigated changes in classification accuracy while identifying five individual tree species (Douglas-fir, western redcedar, bigleaf maple, red alder, and black cottonwood) in the Pacific Northwest United States using two data sets: discrete point Lidar data alone and discrete point data in combination with waveform Lidar data. Waveform information included variables which summarize the frequency domain representation of all waveforms crossing individual trees. Discrete point data alone provided 79.2 percent overall accuracy (kappa = 0.74) for all 5 species and up to 97.8 percent (kappa = 0.96) when comparing individual pairs of these 5 species. Incorporating waveform information improved the overall accuracy to 85.4 percent (kappa = 0.817) for five species, and in several two-species comparisons. Improvements were most notable in comparing the two conifer species and in comparing two of the three hardwood species.

Monitoring Urban Tree Cover Using Object-Based Image Analysis and Public Domain Remotely Sensed Data

Urban forest ecosystems provide a range of social and ecological services, but due to the heterogeneity of these canopies their spatial extent is difficult to quantify and monitor. Traditional per-pixel classification methods have been used to map urban canopies, however, such techniques are not generally appropriate for assessing these highly variable landscapes. Landsat imagery has historically been used for per-pixel driven land use/land cover (LULC) classifications, but the spatial resolution limits our ability to map small urban features. In such cases, hyperspatial resolution imagery such as aerial or satellite imagery with a resolution of 1 meter or below is preferred. Object-based image analysis (OBIA) allows for use of additional variables such as texture, shape, context, and other cognitive information provided by the image analyst to segment and classify image features, and thus, improve classifications. As part of this research we created LULC classifications for a pilot study area in Seattle, WA, USA, using OBIA techniques and freely available public aerial photography. We analyzed the differences in accuracies which can be achieved with OBIA using multispectral and true-color imagery. We also compared our results to a satellite based OBIA LULC and discussed the implications of per-pixel driven vs. OBIA-driven field sampling campaigns. We demonstrated that the OBIA approach can generate good and repeatable LULC classifications suitable for tree cover assessment in urban areas. Another important finding is that spectral content appeared to be more important than spatial detail of hyperspatial data when it comes to an OBIA-driven LULC.

Estimation of forest structure and canopy fuel parameters from small-footprint full-waveform LiDAR data

Precise knowledge of fuel conditions is important for predicting fire hazards and simulating fire growth and intensity across the landscape. We present a methodology to retrieve and map forest canopy fuel and other forest structural parameters using small-footprint full-waveform airborne light detection and ranging (LiDAR) data. Full-waveform LiDAR sensors register the complete returned backscattered signal through time and can describe physical properties of the intercepted objects. This study was undertaken in a mixed forest dominated by Douglas-fir, occasionally mixed with other conifers, in north-west Oregon (United States). We extracted two sets of LiDAR metrics using pulse detection and waveform modelling and then constructed several predictive models using forward stepwise multiple linear regression. The resulting models explained ~80% of the variability for many of the canopy fuel and forest structure parameters: aboveground biomass (R2=0.84), quadratic mean diameter (R2=0.82), canopy height (R2=0.79), canopy base height (R2=0.78) and canopy fuel load (R2=0.79). The lowest performing models included basal area (R2=0.76), stand volume (R2=0.73), canopy bulk density (R2=0.67) and stand density index (R2=0.66). Our results indicate that full-waveform LiDAR systems show promise in systematically characterising the structure and canopy fuel loads of forests, which may enable accurate fire behaviour forecasting that in turn supports the development of prevention and planning policies.

Object-based classification of semi-arid wetlands

Wetlands are valuable ecosystems that benefit society. However, throughout history wetlands have been converted to other land uses. For this reason, timely wetland maps are necessary for developing strategies to protect wetland habitat. The goal of this research was to develop a time-efficient, automated, low-cost method to map wetlands in a semi-arid landscape that could be scaled up for use at a county or state level, and could lay the groundwork for expanding to forested areas. Therefore, it was critical that the research project contain two components: accurate automated feature extraction and the use of low-cost imagery. For that reason, we tested the effectiveness of geographic object-based image analysis (GEOBIA) to delineate and classify wetlands using freely available true color aerial photographs provided through the National Agriculture Inventory Program. The GEOBIA method produced an overall accuracy of 89% (khat = 0.81), despite the absence of infrared spectral data. GEOBIA provides the automation that can save significant resources when scaled up while still providing sufficient spatial resolution and accuracy to be useful to state and local resource managers and policymakers.

Hyperspectral Analysis of Soil Nitrogen, Carbon, Carbonate, and Organic Matter Using Regression Trees

The characterization of soil attributes using hyperspectral sensors has revealed patterns in soil spectra that are known to respond to mineral composition, organic matter, soil moisture and particle size distribution. Soil samples from different soil horizons of replicated soil series from sites located within Washington and Oregon were analyzed with the FieldSpec Spectroradiometer to measure their spectral signatures across the electromagnetic range of 400 to 1,000 nm. Similarity rankings of individual soil samples reveal differences between replicate series as well as samples within the same replicate series. Using classification and regression tree statistical methods, regression trees were fitted to each spectral response using concentrations of nitrogen, carbon, carbonate and organic matter as the response variables. Statistics resulting from fitted trees were: nitrogen R2 0.91 (p < 0.01) at 403, 470, 687, and 846 nm spectral band widths, carbonate R2 0.95 (p < 0.01) at 531 and 898 nm band widths, total carbon R2 0.93 (p < 0.01) at 400, 409, 441 and 907 nm band widths, and organic matter R2 0.98 (p < 0.01) at 300, 400, 441, 832 and 907 nm band widths. Use of the 400 to 1,000 nm electromagnetic range utilizing regression trees provided a powerful, rapid and inexpensive method for assessing nitrogen, carbon, carbonate and organic matter for upper soil horizons in a nondestructive method.

Improved Salient Feature-Based Approach for Automatically Separating Photosynthetic and Nonphotosynthetic Components Within Terrestrial Lidar Point Cloud Data of Forest Canopies

Accurate separation of photosynthetic and nonphotosynthetic components in a forest canopy from 3-D terrestrial laser scanning (TLS) data is a challenging but of key importance to understand the spatial distribution of the radiation regime, photosynthetic processes, and carbon and water exchanges of the forest canopy. The objective of this paper was to improve current methods for separating photosynthetic and nonphotosynthetic components in TLS data of forest canopies by adding two additional filters only based on its geometric information. By comparing the proposed approach with the eigenvalues plus color information-based method, we found that the proposed approach could effectively improve the overall producers accuracy from 62.12% to 95.45%, and the overall classification producers accuracy would increase from 84.28% to 97.80% as the forest leaf area index (LAI) decreases from 4.15 to 3.13. In addition, variations in tree species had negligible effects on the final classification accuracy, as shown by the overall producers accuracy for coniferous (93.09%) and broadleaf (94.96%) trees. To remove quantitatively the effects of the woody materials in a forest canopy for improving TLS-based LAI estimates, we also computed the “woody-to-total area ratio” based on the classified linear class points from an individual tree. Automatic classification of the forest point cloud data set will facilitate the application of TLS on retrieving 3-D forest canopy structural parameters, including LAI and leaf and woody area ratios.

Spatial variability of terrestrial laser scanning based leaf area index

Abstract Forest stand point clouds generated from multiple scan locations using terrestrial laser scanning (TLS) have diverse range of spatial distribution patterns. These in turn have an effect on the direct leaf area index (LAI) estimation from the point cloud. However, the most effective placement of the scanning equipment in homogeneous vs. heterogeneous stands has not been investigated. In this research, TLS was used to sample an evenly planted Douglas-fir (Pseudotsuga menziesii) seedling forest stand and a mature heterogeneous forest stand dominated by Douglas-fir (P. menziesii) and Western hemlock (Tsuga heterophylla). A new method, circular point cloud slicing, was developed to explore the spatial variation of point density for both azimuthal angular and radial directions. The results show that alone, a central location 360° scan data, does not capture all of the stand characteristics and less than 50% of variation of the estimation of effective leaf area index (LAIe) of a mature heterogeneous stand. Thus, reducing occlusion, by incorporating additional lateral side view scans, is necessary to comprehensively represent the canopy structure, and structural variation of the heterogeneous forest stand. It was also shown, based on the assumption that the comprehensive scan combination can fully represent the forest stand, and that LAIe estimated from the comprehensive multi-direction mosaiced dataset are higher by twofold compared to the result from central scan only.

Fourier transformation of waveform Lidar for species recognition

In precision forestry, tree species identification is one of the critical variables of forest inventory. Lidar, specifically full-waveform Lidar, holds high promise in the ability to identify dominant hardwood tree species in forests. Raw waveform Lidar data contain more information than can be represented by a limited series of fitted peaks. Here we attempt to use this information with a simple transformation of the raw waveform data into the frequency domain using a fast Fourier transform. Some relationships are found among the influences of component frequencies across a given species. These relationships are exploited using a classification tree approach to separate three hardwood tree species native to the Pacific Northwest of the United States. We are able to correctly classify 75% of the trees ( 0.615) in our training data set. Each trees species was predicted using a classification tree built from all the other training trees. Two of the species grow in proximity and grow to a similar form, making differentiation difficult. Across all the classification trees built during the analysis, a small group of frequencies is predominantly used as predictors to separate the species.