• Chivenge, Emmerson
  University of Maine,
  

One of the robust and measures of stand density is Reineke’s (1933) stand density index (SDI), which has been used for predicting stand development and self-thinning in single-species, even-aged stands and stand density management diagrams (SDMDs). Thus, the SDI has since been modified for application in multi-cohort and mixed composition stands to understand density-growth dynamics and guiding forest management operations. The dissertation synthesized literature towards the modifications, statistical methods and necessary data for estimating SDI and application of the original SDI in multi-cohort and mixed stands. The modified SDI has been applied in multi-cohort, mixed composition stands using robust statistical methods such as hierarchical Bayesian methods and linear quantile mixed modeling.

There has been a shift towards permanent plot data and repeated measurements from national inventories.  The robust statistical methods incorporate ancillary data such as climate information and functional traits for example wood specific gravity, drought, and shade tolerance. Forest Inventory and Analysis (FIA) is responsible for providing characteristics and statistics regarding forest attributes and  ecosystem processes at various strategic scales each with their own levels of precision and refinement. However, these characteristics and attributes are available at very coarse resolution across the CONUS. The study leveraged a developed raster (TREEMAP; Scientific Data 8, 11) to produce map and estimate SDI, maximum SDI (SDIMAX) and relative density at 30 x 30m resolution. The differences in FIA and TREEMAP based estimates where highest at low spatial scale such as counties. The differences in the estimates can be attributed to differences in the spatial resolution, underlying assumptions of each method, and spatial-temporal misalignment between the estimates.

Based on the modification and applications done to the original SDI, there was need to empirically test the relationship between RD and different growth metrics. Gross growth, net growth and mortality linearly increased with increases in RD before reaching the breakpoint. Specifically, across, ecological subsections, gross and net growth increased up to breakpoints of RD=0.51 and 0.52 respectively forest types and ecosubsections. Basing on the right slope coefficients, the relationships between RD, gross and net growth and mortality generally remained constant after the breakpoint. Stand structure and density identified as the potential drivers of the gross and net growth. Stand structure was the main driver for mortality.

Overall, this study synthesized literature on the refinement and application of Reineke’s SDI in temperate forests, how national inventory data can be integrated with Geographical Information Systems and Remote Sensing to rapidly map and estimate stand density metrics at very low resolution and empirically derive a change point in the RD growth relationship. The findings inform policy development at the regional or subnational, national scales and management decisions at the project scale.