From the April 1994 issue of "Forest Research West"

Predicting Young Conifer Stand Dynamics


by J. Louise Mastrantonio

The evolution of growth and yield information in the United States has resulted in a transition from the traditional presentation of yield tables to computer-based stand simulators. At the Pacific Southwest Research Station's Redding Silviculture Laboratory, work is underway to take another step in this evolutionary process. A model called CONIFERS is under development that will employ the power of object-oriented programming to predict the dynamics of forest ecosystems in early successional stages in northern California and southern Oregon.

Little known and understood outside the ranks of professional forestry, growth and yield prediction systems are nevertheless the building blocks of modern forest management. With them, foresters can predict tree growth and wood volume production. Such systems have traditionally been used to develop annual timber harvest levels (allowable cuts), plan for sustained yield, and as an aid in evaluating the economic benefits of various cultural practices: density management, control of competing vegetation, and fertilization. Without them, timber harvest prediction would be little more than an educated guess.

Much of the contemporary work in growth and yield in this country is based upon the foundation built in the 1920's and 1930' by researchers such as Donald Bruce, Richard McArdie, Walter Meyer, Duncan Dunning and L. H.. Reinecke. Early work in this field featured normal yield tables. Such tables have allowed forest managers to estimate future yields of fully stocked (usually single species) stands, as a function of site index and stand age.

In the late 1960's, growth and yield research moved forward with the development of computer-based simulation systems. One of the most notable example of this development was the release of the Prognosis model for the intermountain region. The Prognosis model allowed users to make predictions of mixed species stands, and produce much more detailed output than could be obtained from traditional yield tables.

Currently there are four computer growth models available for use in northern California and southwest Oregon.

These models all run on personal computers (Prognosis was originally developed for main-frame computers but has recently been adapted for the PC). With the exception of SYSTUM-1, these models are used to predict growth and yield of established stands in which shrubs are not influencing tree growth. While these simulators work well for many stands, they share a significant shortcoming. They are not well suited to young stands, particular young plantations , growing with a significant component of non-tree vegetation.

SYSTUM-1 is a model built for predicting the growth of young plantations. The SYSTUM-1 project was one of the first to attempt to address the problems of predicting the growth of young stands of trees and shrubs. Two problems revealed by the SYSTUM-1 project were the difficulties of predicting the growth of non-tree vegetation and the limited utility of the traditional approaches to forestry software development and interface design.

Conifers

CONIFERS (California and Oregon Interactive Forest Ecosystem Response Simulator) is a model being developed by Martin Ritchie of the Pacific Southwest Research Station in Redding, California. This model continues the work on young stands started with the SYSTUM-1 modeling project. There are, however, a number of substantial differences between the two. The primary differences are in the way non-tree vegetation is simulated and in the structuring of the software itself.

The idea behind CONIFERS is to simulate young stand growth. "You input a list of sample trees from a plantation and the simulator attempts to predict how that stand will develop over time," according to Ritchie.

In CONIFERS, the tree (or plant) is the basic modeling unit and height growth is the primary driving variable. In addition to height growth, the model also predicts changes in crown size, tree diameter and probability of mortality.

CONIFERS requires three types of data as input:

  1. Plant -level data -- a sample list of trees (and, optionally, shrubs) from the target stand. These must be from a statistically valid sample of plots within the stand being studied. The primary required observations include a plot identifier, species code, height and expansion factor. Other variables in the standard file are current annual height increment, crown width, crown ratio, diameter at breast height, and diameter at 10 centimeters.
  2. Plot level data includes optional information on percent cover and height, slope and aspect, elevation and universal transverse mercator projection.
  3. Stand-level data include site index and stand age.
Tree- and plant-level information may be input in one of five formats: CACTOS, ORGANON, SYSTUM-1, comprehensive CONIFERS file, or the US Forest Service Region-5 field data recorder file.

In addition to simulating a number of different thinning regimes, CONIFERS is capable of simulating treatment of competing vegetation. In northern California and southwest Oregon, young mixed-conifer stands are often significantly impacted by on-tree vegetation. Often the competition is provided by brush such as ceanothus and manzanita. Thus, the effect of competing vegetation on tree growth is an important factor in predicting stand dynamics.

Ritchie is attempting to develop a more robust methodology for predicting growth of non-tree vegetation. He notes, "We are developing individual plant-based predictors of growth for competing vegetation in an attempt to overcome some of the problems encountered with the SYSTUM-1 model. This is something that has not been done before, to my knowledge."

One other unique aspect of this simulator is the application of object oriented programming in simulator development. The end result of this work is a group of reusable objects in the form of a class-library. The program is being written in the C++ language by programmer Jeff Hamann. Ritchie hopes this will allow for much faster development of software for forestry applications. "We have been reinventing the wheel with respect to software development, and this new direction will provide a general forest stand analysis application class-library. We have already used this class-library in a number of other applications which have nothing to do with simulating growth and yield. It is a very powerful tool," he says.

Ritchie describes CONIFERS as an "interactive system that is more user-friendly than earlier growth and yield simulators." It enables managers to analyze the pros and cons of various management alternatives; that is, to test various management schemes before they are applied on the ground. The model has the appearance and feel of many contemporary software products, with drag and drop menus, mouse support and context-sensitive help.

The CONIFERS simulator also provides flexible, user defined output variables that are established "on the fly." This is important because forest management goals often change over time, and resource managers must constantly respond to meet new needs and challenges presented by society. This flexible output may take the form of a user-designed table, a spreadsheet output file for graphical analysis, or a plot summary file including spacial data when available.

New developments are just around the corner for this effort. "We want to get away from focusing on classic plantation forestry and take better advantage of the capabilities of modern computer systems," Ritchie says. Wood volume production is by no means the only application of such simulators. Another application of models like CONIFERS is in evaluating wildlife habitat. Models can be designed to estimate stand characteristics that can be related to habitat suitability. Examples include snag retention, and more specific descriptors of stand structure and non-tree vegetation. By including flexible, on-the-fly, user-defined output, it is hoped that the model will allow users to address a variety of issues. For example, the spacial arrangement of some stand parameters may be analyzed by means of a plot summary file. This will allow the user to evaluate such things as the spacial distribution of regeneration in an area.

The model has not been released to cooperators, but a first release may be ready soon. The model is being beta-tested at this point. Currently, context-sensitive help is being refined for the simulator, so that all needed documentation will be available on screen.

Cooperators

CONIFERS is a cooperative effort of the Pacific Southwest Research Station and the California Forest Research Association (CFRA). The CFRA is a forestry research support group that includes public and private forest resource management concerns. The work was done by scientists at the Silviculture Laboratory of the Forest Service's Pacific Southwest Research Station at Redding, California, under the auspices of the CFRA's Small Trees Modeling Project. The CFRA is comprised of private and public forest resource managers.

For those interested in the CONIFERS model, Ritchie will be developing a user's guide for the program. For further information about other models mentioned in this article, contact the Southwest research Station and request the following:

User's Guide for SYSTUM-1 (Version 2.0): A Simulator of Growth Trends in Young Stands Under Management in California and Oregon, General Technical Report PSW-147, by M.W. Ritchie and R.F. Powers.

Predicting Height Increment of Young-Growth Red Fir in California and Southern Oregon, Research Paper PSW-214, by K.L. Dolph.