Estimating forest attributes using laser scanning data and dual-band, single-pass interferometric aperture radar to improve forest management
MetadataShow full item record
The overall objectives of this dissertation were to (1) determine whether leaf area index (LAI) (Chapter 2), as well as stem density and height to live crown (Chapter 3) can be estimated accurately in intensively managed pine plantations using small-footprint, multiple-return airborne laser scanner (lidar) data, and (2) ascertain whether leaf area index in temperate mixed forests is best estimated using multiple-return airborne laser scanning (lidar) data or dual-band, single-pass interferometric synthetic aperture radar data (from GeoSAR) alone or both in combination (Chapter 4). In situ measurements of LAI, mean height, height to live crown, and stem density were made on 109 (LAI) or 110 plots (all other variables) under a variety of stand conditions. Lidar distributional metrics were calculated for each plot as a whole as well as for crown density slices (newly introduced in this dissertation). These metrics were used as independent variables in best subsets regressions with LAI, number of trees, mean height to live crown, and mean height (measured in situ) as the dependent variables. The best resulting model for LAI in pine plantations had an R2 of 0.83 and a cross-validation (CV) RMSE of 0.5. The CV-RMSE for estimating number of trees on all 110 plots was 11.8 with an R2 of 0.92. Mean height to live crown was also well-predicted (R2 = 0.96, CV-RMSE = 0.8 m) with a one-variable model. In situ measurements of temperate mixed forest LAI were made on 61 plots (21 hardwood, 36 pine, 4 mixed pine hardwood). GeoSAR metrics were calculated from the X-band backscatter coefficients (four looks) as well as both X- and P-band interferometric heights and magnitudes. Both lidar and GeoSAR metrics were used as independent variables in best subsets regressions with LAI (measured in situ) as the dependent variable. Lidar metrics alone explained 69% of the variability in temperate mixed forest LAI, while GeoSAR metrics alone explained 52%. However, combining the LAI and GeoSAR metrics increased the R2 to 0.77 with a CV-RMSE of 0.42. Analysis of data from active sensors shows strong potential for eventual operational estimation of biophysical parameters essential to silviculture.
- Doctoral Dissertations