A Survival Model for Shortleaf Pine Tress Growing in Uneven-Aged Stands

A survival model for shortleaf pine (Pinus echinata Mill.) trees growing in uneven-aged stands was developed using data from permanently established plots maintained by an industrial forestry company in western Arkansas. Parameters were fitted to a logistic regression model with a Bernoulli dependent variable in which ‘1” represented individual tree survival and “0” represented individual tree mortality. Predictions from the model can be interpreted as probabilities of survival. The most important independent variable for prediction of survival probability was the ratio of quadratic mean stand d.b.h. to tree d.b.h. The data were used to evaluate the performance of the model by d.b.h. classes. The model was developed for use in an individual-tree growth simulator for uneven-aged shortleaf pine forests. I N T R O D U C T I O N Despite the economic importance and wide distribution of shortleaf pine relatively little effort has been directed at modeling individual tree survival. Lynch and others [in press(a)] developed a model for individual tree shortleaf pine survival in even-aged forest stands. Individual tree level equations for shortleaf pine dynamics are part of Central States Twigs (Miner and others 1989) and a multipurpose forest projection system for southern forests developed by Boulton and Meldahl (1990). Apparently, no survival models have previously been developed specifically for shottleaf pine managed in uneven-aged stands, though Murphy and Shelton (1996) developed a survival model for individual trees in uneven-aged Joblolly pine (Pinus taeda L.) stands. This paper will present results from the development of a model for survival of individual shortleaf pine trees growing in fo res t s tands under uneven-aged management . Shortleaf pine has a greater natural range than any of the other southern pines and is second only to loblolly in volume (Willet 1986) The species is especially important in the Ouachita mountain region of western Arkansas and eastern Oklahoma. Forecasts of stand dynamics for uneven-aged shortleaf pine stands are important on many acres of forest land managed by public agencies, non-industrial private owners, and certain forest industries in western Arkansas and eas te rn Ok lahoma. Though the uneven-aged sys tem produces somewhat lower merchantab le vo lume than evenaged management it has traditionally been utilized by certain forest industries in the West-Gulf region to produce dimension lumber (Guldin and Baker 1988). Attractive features of the system include low-cost regeneration and relatively high sawtimber volume growth. These features make uneven-aged management of southern pine a viable alternative, especially on lower-quality sites (Guldin and Baker 1988, Shelton and Murphy 1994). A discussion of selection management for shortleaf pine in the Ouachita mountains has been given by Murphy and others (1991). Baker and others (1996) have elucidated the principles of uneven-aged management for loblolly and shortleaf pine. Most of the information currently available for stand dynamics of shortleaf pine growing in naturally regenerated stands is based on data from even-aged stands. USDA Miscellaneous Publication 50 (USDA Forest Serv. 1929) includes normal yield tables for shortleaf pine which were based on data obtained from fully-stocked temporary plots. Yield tables developed by Schumacher and Coile (1960) were developed from 74 “well-stocked” temporary plots. Murphy and Beltz (1981) and Murphy (1982) developed growth and yield equations for shortleaf pine based on Forest Inventory and Analysis plots, most of which were located in unmanaged forests. Murphy (1986) gives a comprehensive account of the growth and yield information available for shortleaf pine prior to 1986. Lynch and others 1999b and Huebschmann and others (1998) describe the Shortleaf Pine Stand Simulator (SLPSS), an individual tree model for even-aged shortleaf pine stands which contains a prediction equation for probability of tree survival. This model is based on remeasured plots located in the Ozark and Ouachita National Forests and distributed over a range of ages, densities and site qualities. Murphy and Farrar (1985) have developed equat ions describing the growth and yield of uneven-aged shortleaf pine stands. These equations describe growth and yield on a stand-level basis. Since the equations describe net yields, there are no explicit predictions for survival or mortality. Murphy and Farrar (1988) proposed a framework for growth and y ie ld model deve lopment in uneven-aged lob lo l ly shortleaf stands which used the Weibull distribution to predict tree diameter distributions. This framework also consisted of stand-level equations.