# Predicting Stem Total and Assortment Volumes in an Industrial Pinus taeda L. Forest Plantation Using Airborne Laser Scanning Data and Random Forest

^{1}

^{2}

^{3}

^{4}

^{5}

^{6}

^{7}

^{*}

## Abstract

**:**

^{2}equal to 0.98, 0.98 and 0.96, with unbiased predictions of −0.17%, −0.12% and −0.23%, and Root Mean Square Error (RMSE) values of 7.83%, 7.71% and 8.63%. Our methodology makes use of commercially available airborne lidar and widely used mathematical tools to provide solutions for increasing the industry efficiency in monitoring and managing wood volume.

## 1. Introduction

^{3}·ha

^{−1}·year

^{−1}[2]. Moreover, P. taeda is commonly managed for production of multiple types of wood such as stem total, saw logs, pulpwood and small-diameter logs and branches, which are used for energy. Saw logs and pulpwood can be further divided into different assortments that differ in size and therefore in economic value [3].

## 2. Methods

#### 2.1. Study Area Description

^{−1}, respectively. The climate of the region is characterized as warm and temperate [28], with annual average precipitation of approximately 1378 mm and an annual average temperature of 18.4 °C. The P. taeda stands are situated on a plateau where the topography is relatively flat. The plantations are managed by Klabin S.A., a pulp and paper company.

#### 2.2. Field Data Collection

^{2}(i.e., 20 m × 30 m) were randomly established and measured across 50 stands distributed in four plantations. As such, the sample plots well represent the study area, and they capture the entire structural variability in these stands with ages ranging from three to nine years old. All plots were geo-referenced with a geodetic GPS with differential correction capability (Trimble Pro-XR, Trimble, Sunnyvale, CA, USA) ensuring a location error lower than 10 cm. In each sample plot, individual trees were measured for dbh (diameter at breast height) at 1.30 m and a random subsample (15%) of trees for tree height (Ht). For those trees in the plots that were not directly measured for Ht, the inventory team of Klabin S.A. predicted heights from hypsometric equations [29], employing dbh as the independent variable, and Ht as the dependent variable, following the model below:

^{2}) and standard error of estimate in percentage (SEE%) of the models ranged from 0.96 to 0.98 and 5.1 to 6.5, respectively.

_{i}= stem diameter (cm) at the ith position; dbh = diameter (cm) at breast height (1.30 m); h = total height (m); h

_{i}= height (m) at the ith position; and K = π/40,000 is an adjustment factor to estimate volume as m

^{3}·ha

^{−1}.

^{2}and standard error of the estimate (SEE; given in %) for the polynomial models used in this study are presented in Table 1.

^{3}·ha

^{−1}for each class of stand ages is presented in Table 2. SEE (%) is the standard error of the estimate, expressed as a percentage.

#### 2.3. Lidar Data Acquisition and Data Processing

#### 2.4. Predictor Variables Selection

#### 2.5. Random Forest Model Development

^{2}, Root Mean Square Error (RMSE), and Bias (both absolute and relative) by the linear relationship between predicted (output from RF) and observed stem volumes:

_{i}is the observed value for plot i, and ${\widehat{y}}_{i}$ is the predicted value for plot i. Moreover, relative RMSE (%) and biases (%) were calculated by dividing the absolute values (Equations (5) and (6)) by the mean of the observed stem volume. Based on earlier experiences and recommendations from literature [4,5], we defined acceptable model accuracy as a relative RMSE and Bias of <15%.

^{2}, absolute and relative RMSE and Bias were computed based on the linear relationship between observed and predicted volumes using the holdout samples. We used also two-sided Kolmogorov-Smirnov (KS) in R [38] and a statistical equivalence test [41] to compare the field- and lidar-based stem volume estimates in each iteration.

#### 2.6. Predictive Stem Volumes Maps

## 3. Results

#### 3.1. Predictor Variable Selection

#### 3.2. Model Performances

^{2}, RMSE and Bias for all 500 bootstrap runs (Table 8). Observed and predicted stem volumes in each bootstrap iteration did not differ significantly by the statistical KS and equivalence test (p-values > 0.05) as well. Overall, all models using H99TH and HSKEW performed very well, with relative RMSE and Bias <15% in the bootstrap procedure. The observed and the average of the predicted stem volumes from the 500 bootstrap runs were also compared and according to the KS and equivalence tests those values did not differ significantly (p-values > 0.05) too (Figure 5).

#### 3.3. Prediction Maps

## 4. Discussion

^{2}of ~0.65 for estimating both saw log and pulpwood volumes. While those authors have showed the great potential of lidar in retrieving assortment volumes, this specific application is still relatively novel and further studies, such as presented herein, still need to be carried out.

## 5. Conclusions

## Acknowledgments

## Author Contributions

## Conflicts of Interest

## References

- Payn, T.; Carnus, J.-M.; Freer-Smith, P.; Kimberley, M.; Kollert, W.; Liu, S.; Orazio, C.; Rodriguez, L.; Silva, L.N.; Wingfield, M.J. Changes in planted forests and future global implications. For. Ecol. Manag.
**2015**, 352, 57–67. [Google Scholar] [CrossRef] - Indústria Brasileira de Árvores (IBÁ). Brazilian Tree Industry. 2015. Available online: http://www.iba.org/images/shared/iba_2015.pdf (accessed on 10 November 2016).
- Kohler, S.V.; Wolff, N.I.; Figueiredo Filho, A.; Arce, J.E. Dynamic of assortment of Pinus taeda L. plantation in different site classes in Southern Brazil. Sci. For.
**2014**, 42, 403–410. [Google Scholar] - Silva, C.A.; Klauberg, C.; Hudak, A.T.; Vierling, L.A.; Liesenberg, V.; Bernett, L.G.; Scheraiber, C.F.; Schoeninger, E.R. Estimating Stand Height and Tree Density in Pinus taeda plantations using in-situ data, airborne LiDAR and k-Nearest Neighbor Ismputation. Ann. Braz. Acad. Sci.
**2017**, 90, 1–15, in press. [Google Scholar] - Silva, C.A.; Klauberg, C.; Hudak, A.T.; Vierling, L.A.; Liesenberg, V.; Carvalho, S.P.; Rodriguez, L.C. A principal component approach for predicting the stem volume in Eucalyptus plantations in Brazil using airborne Lidar data. Forestry
**2016**, 89, 422–433. [Google Scholar] [CrossRef] - Næsset, E. Determination of mean tree height of forest stands using airborne laser scanner data. ISPRS J. Photogramm.
**1997**, 52, 49–56. [Google Scholar] [CrossRef] - Næsset, E. Predicting forest stand characteristics with airborne scanning laser using a practical two-stage procedure and field data. Remote Sens. Environ.
**2002**, 80, 88–99. [Google Scholar] [CrossRef] - Næsset, E.; Gobakken, T.; Holmgren, J.; Hyyppä, H.; Hyyppä, J.; Maltamo, M.; Nilsson, M.; Olsson, H.; Persson, Å.; Söderman, U. Laser scanning of forest resources: The Nordic experience. Scand. J. For. Res.
**2004**, 19, 482–499. [Google Scholar] [CrossRef] - Næsset, E. Airborne laser scanning as a method in operational forest inventory: Status of accuracy assessments accomplished in Scandinavia. Scand. J. For. Res.
**2007**, 22, 433–442. [Google Scholar] [CrossRef] - Hudak, A.T.; Evans, J.S.; Stuart, A.M. Lidar utility for natural resource managers. Remote Sens.
**2009**, 1, 934–951. [Google Scholar] [CrossRef] - Andersen, H.E.; McGaughey, R.J.; Reutebuch, S.E. Estimating forest canopy fuel parameters using lidar data. Remote Sens. Environ.
**2005**, 94, 441–449. [Google Scholar] [CrossRef] - Hudak, A.T.; Crookston, N.L.; Evans, J.S.; Falkowski, M.J.; Smith, A.M.; Gessler, P.E.; Morgan, P. Regression modeling and mapping of coniferous forest basal area and tree density from discrete-return lidar and multispectral satellite data. Can. J. Remote Sens.
**2006**, 32, 126–138. [Google Scholar] [CrossRef] - White, J.C.; Wulder, M.A.; Buckmaster, G. Validating estimates of merchantable volume from airborne laser scanning (ALS) data using weight scaled data. For. Chron.
**2014**, 90, 378–385. [Google Scholar] [CrossRef] - Silva, C.A.; Klauberg, C.; Carvalho, S.D.P.C.; Hudak, A.T. Mapping aboveground carbon stocks using Liar data in Eucalyptus spp. plantations in the state of São Paulo, Brazil. Sci. For.
**2014**, 42, 591–604. [Google Scholar] - Korhonen, L.; Peuhkurinen, J.; Jukka, M.; Suvanto, A.; Maltamo, M.; Packalen, P.; Kangas, J. The use of airborne laser scanning to estimate sawlog volumes. Forestry
**2008**, 81, 499–510. [Google Scholar] [CrossRef] - Peuhkurinen, J.; Maltamo, M.; Malinen, J. Estimating Species-Specific diameter distributions and saw log recoveries of boreal forests from airborne laser scanning data and aerial photographs: A distribution-based approach. Silva Fenn.
**2008**, 42, 600–625. [Google Scholar] [CrossRef] - Sherrill, J.R.; Bullock, B.P.; Mullin, T.J.; McKeand, S.E.; Purnell, R.C. Total and merchantable stem volume equations for mid rotation loblolly pine (Pinus taeda L.). South. J. Appl. For.
**2011**, 35, 105–108. [Google Scholar] - Hudak, A.T.; Crookston, N.L.; Evans, J.S.; Hall, D.E.; Falkowski, M.J. Nearest neighbor imputation of species-level, plot-scale forest structure attributes from lidar data. Remote Sens. Environ.
**2008**, 112, 2232–2245. [Google Scholar] [CrossRef] - Ahmed, O.S.; Franklin, S.E.; Wulder, M.A.; White, J.C. Characterizing stand-level forest canopy cover and height using landsat time series, samples of airborne lidar, and the random forest algorithm. ISPRS J. Photogramm. Remote Sens.
**2015**, 101, 89–101. [Google Scholar] [CrossRef] - Breiman, L. Random forests. Mach. Learn.
**2001**, 45, 5–32. [Google Scholar] [CrossRef] - Grossmann, E.; Ohmann, J.; Kagan, J.; May, H.; Gregory, M. Mapping ecological systems with a random forest model: Tradeoffs between errors and bias. GAP Anal. Bull.
**2010**, 17, 16–22. [Google Scholar] - Naidoo, L.; Cho, M.A.; Mathieu, R.; Asner, G. Classification of savanna tree species, in the Greater Kruger National Park region, by integrating hyperspectral and Lidar data in a Random Forest data mining environment. ISPRS J. Photogramm. Remote Sens.
**2012**, 69, 167–179. [Google Scholar] [CrossRef] - Yu, X.; Hyyppä, J.; Vastaranta, M.; Holopainen, M.; Viitala, R. Predicting individual tree attributes from airborne laser point clouds based on the random forests technique. ISPRS J. Photogramm. Remote Sens.
**2011**, 661, 28–37. [Google Scholar] [CrossRef] - Ko, C.; Sohn, G.; Remmel, T.K.; Miller, J.R. Maximizing the Diversity of Ensemble Random Forests for Tree Genera Classification Using High Density Lidar Data. Remote Sens.
**2016**, 8, 646. [Google Scholar] [CrossRef] - Olden, J.D.; Lawler, J.J.; Poff, N.L. Machine learning methods without tears: A primer for ecologists. Q. Rev. Biol.
**2008**, 83, 171–193. [Google Scholar] [CrossRef] [PubMed] - Stumpf, A.; Kerle, N. Object-oriented mapping of landslides using Random Forests. Remote Sens. Environ.
**2011**, 115, 2564–2577. [Google Scholar] [CrossRef] - Lawrence, R.L.; Wood, S.D.; Sheley, R.L. Mapping invasive plants using hyperspectral imagery and Breiman Cutler classifications (RandomForest). Remote Sens. Environ.
**2006**, 100, 356–362. [Google Scholar] [CrossRef] - Köppen, W.; Geiger, R. Klimakarte der Erde. Wall-Map 150 cm × 200 cm; Verlag Justus Perthes: Gotha, Germany, 1928. [Google Scholar]
- Curtis, R.O. Height-diameter and height-diameter-age equations for second-growth Douglas-fir. For. Sci.
**1967**, 13, 365–375. [Google Scholar] - Schöpfer, W. Automatisierung Des Massem, Sorten Und Wertberechnung Stenender Waldbestande Schriftenreihe Bad; Wurtt-Forstl: Koblenz, Germany, 1966. [Google Scholar]
- McGauchey, R.J. FUSION/LDV: Software for LiDAR Data Analysis and Visualization; Forest Service Pacific Northwest Research Station USDA: Seattle, WA, USA, 2015; Available online: http://forsys.cfr.washington.edu/fusion/ FUSIONmanual.pdf (accessed on 15 October 2015).
- Kraus, K.; Pfeifer, N. Determination of terrain models in wooded areas with airborne laser scanner data. ISPRS J. Photogramm. Remote Sens.
**1998**, 53, 193–203. [Google Scholar] [CrossRef] - Hudak, A.T.; Strand, E.K.; Vierling, L.A.; Byrne, J.C.; Eitel, J.U.H.; Martinuzzi, S.; Falkowski, M.J. Quantifying aboveground forest carbon pools and fluxes from repeat LiDAR surveys. Remote Sens. Environ.
**2012**, 123, 25–40. [Google Scholar] [CrossRef] - Silva, C.A.; Hudak, A.T.; Klauberg, C.; Vierling, L.A.; Gonzalez-Benecke, C.; de Padua Chaves Carvalho, S.; Rodriguez, L.C.E.; Cardil, A. Combined effect of pulse density and grid cell size on predicting and mapping aboveground carbon in fast-growing Eucalyptus forest plantation using airborne LiDAR data. Carbon Balance Manag.
**2017**, 12, 13. [Google Scholar] [CrossRef] [PubMed] - Evans, J.S.; Cushman, S.A. Gradient modeling of conifer species using Random Forests. Landsc. Ecol.
**2009**, 5, 673–683. [Google Scholar] [CrossRef] - Evans, J.S.; Murphy, M.A.; Holden, Z.A.; Cushman, S.A. Modeling species distribution and change using Random Forests. In Predictive Modeling in Landscape Ecology; Drew, C.A., Huettmann, F., Wiersma, Y., Eds.; Springer: New York, NY, USA, 2010; pp. 139–159. [Google Scholar]
- Liaw, A.; Wiener, M. RandomForest: Breiman and Cutler’s Random Forests for Classification and Regression, Version 4.6–12. 2015. Available online: https://cran.rproject.org/web/packages/randomForest/ (accessed on 15 October 2016).
- R Core Team. R: A Language and Environment for Statistical Computing; R Foundation for Statistical Computing: Vienna, Austria, 2017; Available online: http://www.R-project.org (accessed on 20 October 2016).
- Bright, B.C.; Hudak, A.T.; McGaughey, R.; Andersen, H.E.; Negron, J. Predicting live and dead tree basal area of bark beetle affected forests from discrete-return lidar. Can. J. Remote Sens.
**2013**, 39, S99–S111. [Google Scholar] [CrossRef] - Lopatin, J.; Dolos, K.; Hernández, H.J.; Galleguillos, M.; Fassnacht, F.E. Comparing Generalized Linear Models and random forest to model vascular plant species richness using LiDAR data in a natural forest in central Chile. Remote Sens. Environ.
**2016**, 173, 200–210. [Google Scholar] [CrossRef] - Robinson, A.P.; Duursma, R.A.; Marshall, J.D. A regression-based equivalence test for model validation: Shifting the burden of proof. Tree Physiol.
**2005**, 25, 903–913. [Google Scholar] [CrossRef] [PubMed] - Holopainen, M.; Vastaranta, M.; Rasinmäki, J.; Kalliovirta, J.; Mäkinen, A.; Haapanen, R.; Melkas, T.; Yu, X.; Hyyppä, J. Uncertainty in timber assortment estimates predicted from forest inventory data. Eur. J. For. Res.
**2010**, 129, 1131–1142. [Google Scholar] [CrossRef] - Sibona, E.; Vitali, A.; Meloni, F.; Caffo, L.; Dotta, A.; Lingua, E.; Motta, R.; Garbarino, M. Direct Measurement of Tree Height Provides Different Results on the Assessment of LiDAR Accuracy. Forests
**2017**, 8, 7. [Google Scholar] [CrossRef] - Kankare, V.; Vauhkonen, J.; Tanhuanpaa, T.; Holopainen, M.; Vastaranta, M.; Joensuu, M. Accuracy in estimation of timber assortments and stem distribution—A comparison of airborne and terrestrial laser scanning techniques. ISPRS J. Photogramm. Remote Sens.
**2014**, 97, 89–97. [Google Scholar] [CrossRef] - Zhao, K.; Popescu, S.; Meng, X.; Pang, Y.; Agca, M. Characterizing forest canopy structure with lidar composite metrics and machine learning. Remote Sens Environ.
**2011**, 115, 1978–1996. [Google Scholar] [CrossRef] - Mascaro, J.; Asner, G.P.; Knapp, D.E.; Kennedy-Bowdoin, T.; Martin, R.E.; Anderson, C.; Higgins, M.; Chadwick, K.D. A tale of two “forests”: Random forest machine learning AIDS tropical forest carbon mapping. PLoS ONE
**2014**, 9, e85993. [Google Scholar] [CrossRef] [PubMed] - García Gutiérrez, J.; Martínez Álvarez, F.; Troncoso Lora, A.; Riquelme Santos, J.C. A comparison of machine learning regression techniques for LiDAR-derived estimation of forest variables. Neurocomputing
**2015**, 167, 24–31. [Google Scholar] [CrossRef] - Hudak, A.T.; Bright, B.C.; Pokswinski, S.M.; Loudermilk, E.L.; O’Brien, J.J.; Hornsby, B.S.; Klauberg, C.; Silva, C.A. Mapping forest structure and composition from low-density LiDAR for informed forest, fuel, and fire management at Eglin Air Force Base, Florida, USA. Can. J. Remote Sens.
**2016**, 42, 411–427. [Google Scholar] [CrossRef] - Hayashi, R.; Weiskittel, A.; Sader, S. Assessing the feasibility of low-density lidar for stand inventory attribute predictions in complex and managed forests of Northern Maine, USA. Forests
**2014**, 5, 363–383. [Google Scholar] [CrossRef] - Kankare, V.; Vastaranta, M.; Holopainen, M.; Raty, M.; Yu, X.; Hyyppa, J.; Hyyppa, H.; Alho, P.; Viitala, R. Retrieval of forest aboveground biomass and stem volume with airborne scanning LiDAR. Remote Sens.
**2013**, 5, 2257–2274. [Google Scholar] [CrossRef] [Green Version] - Wu, J.; Yao, W.; Choi, S.; Park, T.; Myneni, R.B. A Comparative Study of Predicting DBH and Stem Volume of Individual Trees in a Temperate Forest Using Airborne Waveform LiDAR. IEEE Geosci. Remote Sens. Lett.
**2015**, 12, 2267–2271. [Google Scholar] [CrossRef] - Shataeea, S.; Weinaker, H.; Babanejad, M. Plot-level Forest Volume Estimation Using Airborne Laser Scanner and TM Data, Comparison of Boosting and Random Forest Tree Regression Algorithms. Procedia Environ. Sci.
**2011**, 7, 68–73. [Google Scholar] [CrossRef] - Peuhkurinen, J.; Maltamo, M.; Malinen, J.; Pitkänen, J.; Packalén, P. Pre-harvest measurement of marked stands using airborne laser scanning. For. Sci.
**2007**, 53, 653–661. [Google Scholar] - Holmgren, J.; Barth, A.; Larsson, H.; Olsson, H. Prediction of stem attributes by combining airborne laser scanning and measurements from harvesters. Silva Fen.
**2012**, 46, 227–239. [Google Scholar] [CrossRef] - Hawbaker, T.J.; Gobakken, T.; Lesak, A.; Trømborg, E.; Contrucci, K.; Radeloff, V. Light detection and ranging-based measures of mixed hardwood forest structure. For. Sci.
**2010**, 56, 313–326. [Google Scholar] - Van Aardt, J.A.N.; Wynne, R.H.; Oderwald, R.G. Forest Volume and Biomass Estimation Using Small-Footprint Lidar Distributional Parameters on a Per-Segment Basis. For. Sci.
**2006**, 52, 636–649. [Google Scholar] - Ota, T.; Ahmed, O.S.; Franklin, S.E.; Wulder, M.A.; Kajisa, T.; Mizoue, N. Estimation of Airborne Lidar Derived Tropical Forest Canopy Height Using Landsat Time Series in Cambodia. Remote Sens.
**2014**, 6, 10750–10772. [Google Scholar] [CrossRef] - Coops, N.C.; Hilker, T.; Wulder, M.A.; St-Onge, B.; Newnham, G.; Siggins, A. Estimating canopy structure of Douglas-fir forest stands from discrete-return lidar. Trees
**2007**, 21, 295–310. [Google Scholar] [CrossRef] - Tilley, B.K.; Munn, I.A.; Evans, D.L.; Parker, R.C.; Roberts, S.D. Cost Considerations of Using Lidar for Timber Inventory. 2004. Available online: http://sofew.cfr.msstate.edu/papers/0504tilley.pdf/ (accessed on 21 March 2016).
- Hummel, S.; Hudak, A.T.; Uebler, E.H.; Falkowski, M.J.; Megown, K.A. A comparison of accuracy and cost of LiDAR versus stand exam data for landscape management on the Malheur national forest. J. For.
**2011**, 109, 267–273. [Google Scholar] - Silva, C.A.; Hudak, A.T.; Vierling, L.A.; Loudermilk, E.L.; O’Brien, J.J.; Hiers, J.K.; Jack, S.B.; Gonzalez-Benecke, C.A.; Lee, H.; Falkowski, M.J.; et al. Imputation of individual longleaf pine forest attributes from field and LiDAR data. Can. J. Remote Sens.
**2016**, 42, 554–573. [Google Scholar] [CrossRef]

**Figure 1.**Location of study area in Telêmaco Borba, Paraná, Brazil. The black dotes indicate the location of the Pinus teada stands.

**Figure 2.**Process of forest volume mesurement. (

**A**) Pinus plantation; (

**B**) Timber harvester and (

**C**) Log segmentation for classes of volume mesurements.

**Figure 3.**Procedure for predicting stem total and assortment volumes in an industrial P. taeda forest plantation using airborne laser scanning data and random forest.

**Figure 4.**Distribution of observed (black line) and predicted (red line)

**s**tem volume from RF. The gray histograms is based from field data. (

**A**) Total volume (Vt) (

**B**) Commercial volume (Vc) and (

**C**) Pulpwood volume (Vp).

**Figure 5.**Equivalence plots of the observed and the mean of predicted Vt (

**A**), Vc (

**B**) and Vp (

**C**) obtained from the 500 bootstrapped RF model runs. (N = 50). The equivalence plot design presented herein is an adaptation of the original equivalence plots presented by [41]. The grey polygon represents the ±25% region of equivalence for the intercept, and the green vertical bar represents a 95% of confidence interval for the intercept. The predicted stem volumes from the RF models are equivalent with reference to the intercept and slope since the green bar is completely within the grey polygon. If the grey polygon is lower than the green vertical bar, the predicted stem volumes are negatively biased; and if it is higher than the green vertical bar, the predicted stem volumes are positively biased. Moreover, the grey dashed line represents the ±25% region of equivalence for the slope, the fit line is within the dotted lines and the black vertical bar is within the gray rectangle, indicating that the pairwise measurements are equivalent. A green bar that is wider than the region outlined by the grey dashed lines indicates highly variable predictions. The white dots are the pairwise measurements, and the solid line is a best-fit linear model for the pairwise measurements. The light grey dashed line represented the relationship 1:1. The horizontal red bars represent the standard deviation of the 500 bootstraping predictions.

**Figure 6.**Predicted Vt, Vc and Vp of P. taeda at stand-level for the studied stands. (

**A**) 3–5 years; (

**B**) 5–7 years and (

**C**) 7–9 years. The thick line in the box indicates the median value of the predicted stem volume. Boxes extend from the 25th to the 75th percentile, whiskers extend 1.5 times the length of the interquartile range above and below the 75th and 25th percentiles. The white dote is the mean of the predicted stem volume, and the vertical read lines represents the standard deviation around the mean (Mean ± SD).

**Figure 7.**Predicted Vt (

**A1**–

**C1**), Vc (

**A2**–

**C2**) and Vp (

**A3**–

**C3**) of P. taeda at stand-level obtained from the RF models. Representative stand of early (i.e., 3–5 years) (

**A1**–

**3**), intermediate (i.e., 5–7 years) (

**B1**–

**3**) and advanced-stages of development (i.e., 7–9 years) (

**C1**–

**3**).

**Figure 8.**Coefficient of variation (CV) maps in percentage (%) of Vt (

**A1**–

**C1**), Vc (

**A2**–

**C2**) and Vp (

**A3**–

**C3**) of P. taeda at stand-level obtained from the 500 RF bootstrapped runs. Representative stand of early (i.e., 3–5 years) (

**A1**–

**3**), intermediate (i.e., 5–7 years) (

**B1**–

**3**) and advanced stages of development (i.e., 7–9 years) (

**C1**–

**3**).

DBH (cm) Range | Adj. R^{2} | SEE (%) | |
---|---|---|---|

dbh | Volume | ||

0.0–17.9 | 0.96 | 9.58 | 11.55 |

18.0–29.9 | 0.98 | 7.99 | 9.33 |

30.0–70.0 | 0.98 | 7.52 | 8.21 |

Ages (I) | Stem Total and Assortment Volumes (m^{3}·ha^{−1}) | N | ||
---|---|---|---|---|

Vt | Vc | Vp | ||

3 ≤ I < 5 | 56.25 ± 10.98 | 47.53 ± 12.15 | 45.67 ± 11.14 | 19 |

5 ≤ I < 7 | 134.20 ± 30.77 | 124.67 ± 30.3 | 114.20 ± 23.41 | 22 |

7 ≤ I < 9 | 169.50 ± 22.86 | 160.2 ± 22.20 | 129.50 ± 24.83 | 13 |

Mean ± Sd | 113.70 ± 52.53 | 103.86 ± 52.99 | 92.13 ± 42.11 | Total = 50 |

Parameter | Value |
---|---|

Scan angle (°) | +/−30° |

Footprint (m) | 0.33 m |

Flight speed (km/h) | 234.0 km/h |

Horizontal accuracy | 10 cm |

Elevation accuracy | 15 cm |

Operating altitude | 666.17 m |

Scan frequency | 300 kHz |

Pulse density | 4 pulse m^{−2} |

**Table 4.**Lidar-derived canopy height metrics considered as candidate variables for predictive V models [31].

Variable | Description |
---|---|

HMIN | Height Minimum |

HMAX | Height Maximum |

HMEAN | Height Mean |

HMAD | Height median absolute deviation |

HSD | Height standard deviation |

HSKEW | Height skewness |

HKURT | Height kurtosis |

HCV | Height coefficient of variation |

HIQ | Height interquartile range |

HMODE | Height mode |

H01TH | Height 1th percentile |

H05TH | Height 5th percentile |

H10TH | Height 10th percentile |

H15TH | Height 15th percentile |

H20TH | Height 20th percentile |

H25TH | Height 25th percentile |

H30TH | Height 30th percentile |

H35TH | Height 35th percentile |

H40TH | Height 40th percentile |

H45TH | Height 45th percentile |

H50TH | Height 50th percentile |

H55TH | Height 55th percentile |

H60TH | Height 60th percentile |

H65TH | Height 65th percentile |

H70TH | Height 70th percentile |

H75TH | Height 75th percentile |

H80TH | Height 80th percentile |

H90TH | Height 90th percentile |

H95TH | Height 95th percentile |

H99TH | Height 99th percentile |

CR | Canopy Relief Ratio ((HMEAN − HMIN)/(HMAX − HMIN)) |

COV | Canopy Cover (Percentage of first return above 1.30 m) |

r | HMIN | HCV | HIQ | HSKEW | HKUR | H99TH | COV |
---|---|---|---|---|---|---|---|

HCV | −0.45 ** | ||||||

HIQ | 0.14 | −0.09 | |||||

HSKEW | −0.30 ** | 0.83 *** | −0.36 * | ||||

HKUR | 0.27 | −0.81 *** | 0.07 | −0.82 *** | |||

H99TH | 0.39 ** | −0.80 *** | 0.61 *** | −0.81 *** | 0.77 *** | ||

COV | 0.23 | −0.74 *** | 0.12 | −0.67 *** | 0.53 *** | 0.58 *** |

**Table 6.**Mean of the model improvement ratio (MIR) among the remained lidar-derived metrics not highly correlated. The bold represents the highest MIR values.

Atributes | LiDAR-Derived Metrics | ||||||
---|---|---|---|---|---|---|---|

HMIN | HCV | HIQ | HSKEW | HKUR | H99TH | COV | |

Vt | 0.16 | 0.40 | 0.18 | 0.75 | 0.31 | 0.99 | 0.12 |

Vc | 0.15 | 0.39 | 0.17 | 0.77 | 0.30 | 0.99 | 0.10 |

Vp | 0.16 | 0.65 | 0.20 | 0.74 | 0.38 | 0.98 | 0.11 |

**Table 7.**Model accuracies of random forest (RF) models per stem volume in terms of Adj. R

^{2}, Root Mean Square Error (RMSE) and bias calculated by the relationship between predicted and observed stem volumes.

Volume | LiDAR Derived Metrics | Adj. R^{2} | RMSE | Bias | ||
---|---|---|---|---|---|---|

m^{3}·ha^{−1} | % | m^{3}·ha^{−1} | % | |||

Vt | 0.97 | 8.91 | 7.83 | −0.19 | −0.17 | |

Vc | H99TH + HSKEW | 0.98 | 8.00 | 7.71 | −0.12 | −0.12 |

Vp | 0.96 | 7.96 | 8.63 | −0.22 | −0.24 |

**Table 8.**Model accuracies per stem volume type. The average and standard deviation of Adj. R

^{2}, RMSE and bias derived from the 500 bootstrap runs are displayed.

Volume | Adj. R^{2} | RMSE | Bias | ||
---|---|---|---|---|---|

m^{3}·ha^{−1} | % | m^{3}·ha^{−1} | % | ||

Vt | 0.94 ± 0.02 | 12.02 ± 2.78 | 9.80 ± 2.18 | −0.58 ± 2.85 | −0.45 ± 2.30 |

Vc | 0.95 ± 0.02 | 11.67 ± 2.76 | 10.31 ± 2.76 | −0.95 ± 2.80 | −0.82 ± 2.45 |

Vp | 0.91 ± 0.04 | 11.83 ± 2.56 | 12.10 ± 2.57 | −0.49 ± 2.73 | −0.54 ± 2.77 |

© 2017 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

## Share and Cite

**MDPI and ACS Style**

Silva, C.A.; Klauberg, C.; Hudak, A.T.; Vierling, L.A.; Jaafar, W.S.W.M.; Mohan, M.; Garcia, M.; Ferraz, A.; Cardil, A.; Saatchi, S.
Predicting Stem Total and Assortment Volumes in an Industrial *Pinus taeda* L. Forest Plantation Using Airborne Laser Scanning Data and Random Forest. *Forests* **2017**, *8*, 254.
https://doi.org/10.3390/f8070254

**AMA Style**

Silva CA, Klauberg C, Hudak AT, Vierling LA, Jaafar WSWM, Mohan M, Garcia M, Ferraz A, Cardil A, Saatchi S.
Predicting Stem Total and Assortment Volumes in an Industrial *Pinus taeda* L. Forest Plantation Using Airborne Laser Scanning Data and Random Forest. *Forests*. 2017; 8(7):254.
https://doi.org/10.3390/f8070254

**Chicago/Turabian Style**

Silva, Carlos Alberto, Carine Klauberg, Andrew Thomas Hudak, Lee Alexander Vierling, Wan Shafrina Wan Mohd Jaafar, Midhun Mohan, Mariano Garcia, António Ferraz, Adrián Cardil, and Sassan Saatchi.
2017. "Predicting Stem Total and Assortment Volumes in an Industrial *Pinus taeda* L. Forest Plantation Using Airborne Laser Scanning Data and Random Forest" *Forests* 8, no. 7: 254.
https://doi.org/10.3390/f8070254