Academia.eduAcademia.edu

Figure 5 – uploaded by Robert A Washington-Allen

See full PDFdownloadDownload figure
TABLE 5 Summary of regression results using normalized height bins NOTES: * Model is significant at a =0.05 Summary and Conclusion

Table 5 Summary of regression results using normalized height bins NOTES: * Model is significant at a =0.05 Summary and Conclusion

Related Figures (8)

FIGURE 1. Location of the Mount Apo Natural Park research site in Mindanao Island, Southern Philippines (Management Zone Map, right, from www.mafi.org.ph)
FIGURE 1. Location of the Mount Apo Natural Park research site in Mindanao Island, Southern Philippines (Management Zone Map, right, from www.mafi.org.ph)
FIGURE 2 Tree height map and stem map of a study plot inside a multiple use zone Regression modeling was employed to relate the above-ground biomass derived using three tropical allometric equations to the height statistics. Two types of regression models were chosen to evaluate the goodness of fit. Generalized linear model (GLM) using maximum likelihood estimation was considered to assess the relationship between these predictor variables. This method estimates the parameters of a statistical model, with the assumption that tree heights are normally (Gaussian) distributed with some unknown mean and variance. The
FIGURE 2 Tree height map and stem map of a study plot inside a multiple use zone Regression modeling was employed to relate the above-ground biomass derived using three tropical allometric equations to the height statistics. Two types of regression models were chosen to evaluate the goodness of fit. Generalized linear model (GLM) using maximum likelihood estimation was considered to assess the relationship between these predictor variables. This method estimates the parameters of a statistical model, with the assumption that tree heights are normally (Gaussian) distributed with some unknown mean and variance. The
FIGURE 3 Data processing flow chart for terrestrial laser scan data To remedy this issue, similar-looking objects (e.g., branches, hanged backpacks) were identified between scans to allow point cloud registration. Initial attempts were made to do tree-level AGB analysis, but an inspection of these point clouds discouraged us from further work on this research objective. Therefore, we directed our efforts to doing plot-level AGB estimation. (Figure 4B). These results may suggest that TLS may not be able to detect point clouds in upper canopies, or those that can be occluded by other under-canopy dense vegetation. This study also shows that the use of terrestrial LiDAR technology can capture accurate understory information.
FIGURE 3 Data processing flow chart for terrestrial laser scan data To remedy this issue, similar-looking objects (e.g., branches, hanged backpacks) were identified between scans to allow point cloud registration. Initial attempts were made to do tree-level AGB analysis, but an inspection of these point clouds discouraged us from further work on this research objective. Therefore, we directed our efforts to doing plot-level AGB estimation. (Figure 4B). These results may suggest that TLS may not be able to detect point clouds in upper canopies, or those that can be occluded by other under-canopy dense vegetation. This study also shows that the use of terrestrial LiDAR technology can capture accurate understory information.
TABLE 1 ~—_ Lake Venado 1 plot biomass using quadratic mean stem diameter (QMSD)
TABLE 1 ~—_ Lake Venado 1 plot biomass using quadratic mean stem diameter (QMSD)
FIGURE 4 Regression model for maximum height (A) and mean height (B) from field and terrestrial LIDAR system (TLS) data NOTES 2 Using the equation from Chave et al. (2005) Biomass of all trees = (Number of trees present in plot x Biomass of tree at midpoint) / 1000
FIGURE 4 Regression model for maximum height (A) and mean height (B) from field and terrestrial LIDAR system (TLS) data NOTES 2 Using the equation from Chave et al. (2005) Biomass of all trees = (Number of trees present in plot x Biomass of tree at midpoint) / 1000
.BLE 3 Plot-level carbon stock estimates by management zone and land cover type corrected (AICc), minimum deviance, minimum Pearson value, or maximum likelihood function can be used to compare model validity. In this case, the model with lower AICc was chosen to depict goodness of fit. AlCc is defined as AIC with a correction factor for finite sample size or when there is a large number of parameter (k). This study considered a small number of observations, with n = 16 and k = 8. AICc is expressed as follows: (Leptospermum polygalifolium Salisb.), are mostly tagged especially in the Lake Venado. There is an ongoing reforestation project in the SPZ and RZ where Tinikaran seedlings are mostly planted. The carbon values differ between the allometries used, where Brown (1997) were two times bigger than Chave et al. (2005).
.BLE 3 Plot-level carbon stock estimates by management zone and land cover type corrected (AICc), minimum deviance, minimum Pearson value, or maximum likelihood function can be used to compare model validity. In this case, the model with lower AICc was chosen to depict goodness of fit. AlCc is defined as AIC with a correction factor for finite sample size or when there is a large number of parameter (k). This study considered a small number of observations, with n = 16 and k = 8. AICc is expressed as follows: (Leptospermum polygalifolium Salisb.), are mostly tagged especially in the Lake Venado. There is an ongoing reforestation project in the SPZ and RZ where Tinikaran seedlings are mostly planted. The carbon values differ between the allometries used, where Brown (1997) were two times bigger than Chave et al. (2005).
TABLE 2 __ Referenced plot biomass from three allometric equations
TABLE 2 __ Referenced plot biomass from three allometric equations
NOTES: * Model is significant ata =0.05 » Model is significant at a =0.10 ABLE 4 Summary of regression results using percentile heights height, as well as percentile heights at the 25th, 50th, 75th, and 95th levels. The final equation for the predictive model is written as follows: linear model which uses stepwise regression to fit the model. The test statistic is an F-test, where information for regression sum of square (SSR) and total sum of square (SST) are given, enabling the computation of R’. The index R’ explains the linear relationship between dependent and independent variables. A variance inflation (VIF) index detects the presence of multicollinearity in the data. A VIF index larger than 10 denotes presence of multicollinearity.
NOTES: * Model is significant ata =0.05 » Model is significant at a =0.10 ABLE 4 Summary of regression results using percentile heights height, as well as percentile heights at the 25th, 50th, 75th, and 95th levels. The final equation for the predictive model is written as follows: linear model which uses stepwise regression to fit the model. The test statistic is an F-test, where information for regression sum of square (SSR) and total sum of square (SST) are given, enabling the computation of R’. The index R’ explains the linear relationship between dependent and independent variables. A variance inflation (VIF) index detects the presence of multicollinearity in the data. A VIF index larger than 10 denotes presence of multicollinearity.
580 California St., Suite 400
San Francisco, CA, 94104
© 2025 Academia. All rights reserved