Figure1. The projected age, density, basal area, and biomass for white oak at 1980 and 2130 for the Central Hardwood Forest Region at 90 m resolution showing decrease in density, increase in age, basal area and biomass, and white oak range expansion.
Figure 2. The projected density and basal area summarized by species for the Central Hardwoos and Central Appalachians under scenarios with and without harvest.
The short- and long-term LANDIS PRO model predictions are evaluated at site and landscape scales. Evaluating site-scale predictions is conducted through comparing predicted results within raster cells with observed values over time in inventory plots randomly sampled across the landscape. Evaluating landscape-scale predictions is conducted through comparing predicted results stratified by environmental drivers (ecoregion or landtype) with observed values in inventory plots aggregated by those drivers. The short-term FLM predictions are evaluated using forest inventory data whereas the long-term FLM predictions are evaluated using stand density management diagrams (SDMDs) and empirical studies.
Figure 3. Measures of goodness-of-fit for evaluating LANDIS PRO short-term predictions (including species density and basal area) against reserved 50% FIA data (validation subset) at 1988, 1993, 2003 and 2008 at site (a,b) and landscape scales (c,d) in Northern Arkansas. Overall, the short-term predictions (1978 to 2008) showed a reasonable level of performance with accuracy better at landscape than the site scales, and there were greater discrepancies in predicted basal area than density.
Figure 4. The representative predicted stand development patterns from 1978 to 2128 plotted on SDMDs including Gingrich (1967) stocking chart (a) and Reineke (1933) density diagrams (b) for a landscape in Northern Arkansas. These diagrams illustrate changes over time for three typical groups that encompassed a wide range of initial stand conditions. Group I represented the development of stands initialized at the stand initiation stage. Group II represented the development of stands initialized at the stem exclusion stage. Group III represented the development of stands initialized at the late-stand initiation stage with high stocking but a small mean diameter. The comparisons between the SDMDs and the established theories of forest stand development suggested that LANDIS PRO predicted reasonable patterns of average stand development for stands representing a wide range of initial conditions.