Click on the title to browse this book

Forest tree ingrowth is a highly variable and largely stochastic process. Consequently, predicting occurrence, frequency, and composition of ingrowth is a challenging task but of great importance in long-term forest growth and yield model projections. However, ingrowth data often require different statistical techniques other than traditional Gaussian regression, because these data are often bounded, skewed, and non-normal and commonly contain a large fraction of zeros. This study presents a set of regression models based on discrete Poisson and negative binomial probability distributions for ingrowth data collected from permanent sample plots in the Acadian Forest Region of North America. Models considered here include regular Poisson, zero-inflated Poisson (ZIP), zero-altered Poisson (ZAP; hurdle Poisson), regular negative binomial (NB), zero-inflated negative binomial (ZINB), and zero-altered negative binomial (ZANB; hurdle NB). Plot-level random effects were incorporated into each of these models. The ZINB model with random effects was found to provide the best fit statistics for modeling annualized occurrence and frequency of ingrowth. The key explanatory variables were stand basal area per hectare, percentage of hardwood basal area, number of trees per hectare, a measure of site quality, and the minimum measured diameter at breast height of each plot. A similar model was developed to predict species composition. All models showed logical behavior despite the high variability observed in the original data.

We sampled dominant and codominant regeneration on 70 clearcuts, 21–35 years old, on the Hoosier National Forest in southern Indiana, USA, to evaluate influence of site variables on the competitive success of natural oak ( Quercus L.) over time. Collected data was compared with data collected on these same sites 20 years prior. Regression tree analysis indicated aspect, natural region, and oak abundance in preharvest stands had the greatest influence on competitive success (relative density, RD) of oak species in the latter sampling, which was then examined across sites as defined by these three variables. Oak RD increased across all mid- and some lower-slope positions, sites on which oaks are expected to be replaced by faster growing species. Drought events between sampling periods apparently contributed to a decline in RD and vigor of yellow-poplar ( Liriodendron tulipifera L.), a major competitor for growing space. Stump sprouts contributed 45% of dominant oak stems. Mean oak diameters were not significantly lower than those of other species groups, with the exception of yellow-poplar in younger stands and at mid-slope positions. Oak species drought tolerance, relative to more mesic species found on these sites, and the large proportion of oak stump sprouts likely contributed to oak competitive success.

Height to live crown base (HCB) is an important input variable for several growth and yield models. Since HCB is rarely measured in the field, it is often predicted using static models. Instead of predicting HCB, the Forest Vegetation Simulator Northeastern Variant (FVS-NE) uses an equation that predicts crown ratio (CR), which has not been well validated. The main goal of the present study was to construct a regional HCB model for thirteen selected tree species of the Acadian Region of North America. The specific objectives were to: 1) evaluate FVS-NE model predictions, 2) compare suitable model forms, and 3) assess influence of various covariates to improve predictions. We evaluated three model forms, namely Holdaway (1986), logistic, and exponential. The findings indicated that FVS-NE models were significantly biased for all species as the overall mean bias and root mean square error (RMSE) were 0.11 m and 1.80 m, respectively. A logistic equation with size (diameter at breast height [DBH], total height [HT] and ratio of DBH to HT), and competition (crown competition factor [CCF] and basal area larger than subject tree [BAL]) gave the best predictions for all species in this analysis. This model had an overall mean bias <0.01 m and an RMSE of 1.59 m, which represents a significant improvement in predictions compared to FVS-NE. Despite the range of species and observed variation in the data, the equations worked well and can be easily calibrated to new stands with a few local observations.