Al Stage
John Bell asked that I attempt to summarize Al Stage's career in forestry research. I approach this task with some trepidation. For the better part of 30 years Al has been my supervisor, my advisor, my co-investigator, and my friend. He has demonstrated an abiding interest in all aspects of forestry - forestry has been his professional and recreational pursuit for over 50 years. Eight years after "retirement," he maintains a near full time research and consulting program. In some respects, Al's dedication to the profession is a reflection of his long time association with Chuck Wellner, who had a tremendous impact on forestry and forest research in the Intermountain Northwest for over 60 years. Al has enormous intellect, great vision, and a depth of experience that is unparalleled in my associations. These attributes, coupled with a quiet assertiveness and a curiously optimistic approach to any problem, have produced enlightened answers to an extraordinarily broad array of forestry research questions. It has been my honor and privilege to work with Al - I hope I can capture some essence of his achievements in this short note.
The Prognosis Model Team from around 1985.
From left: Dave Hamilton, Bill Wykoff, Nick Crookston, Bob Monserud,
Dennis Ferguson, Al Stage, and Melinda Moeur
Al was raised in West Springfield, Massachusetts, and his interest in forestry was likely kindled by his early employment as an assistant to the city's surveyor. In the late 1940's he attended the University of Michigan, and attained a B.S. degree in Forest Management in 1951. He completed a M.F. in Forest Ecology in 1952. In 1951, Al was hired by the Northern Rocky Mountain Forest and Range Experiment Station and, when not on military or educational leave, supervised the Experimental Forest at Priest River until 1956. Al served in the military during the Korean War and leveraged his experience to gain a position as a "computer" with a regiment surveying the Philippines. In 1956, Al moved to Spokane, WA, which was the headquarters of the Inland Empire Research Center, a division of the Intermountain Forest and Range Experiment Station. Contributions during Al's early career included description of a method for calibrating and using the thumb as an angle gauge. He also developed the "Stage Gauge", which is a circular slide rule that greatly facilitated field calculation of horizontal distances, tree heights, and critical distances prior to the advent of portable computing.
In 1959, Al returned to the University of Michigan and earned an M.S. in Mathematical Statistics (1961) and a Ph.D. in Forest Mensuration (1966). His Doctoral research focused on methods of estimating site index for grand fir, a tolerant species that is prone to early suppression. This research introduced two concepts to site index methodology: early suppression can be accounted for by the number of rings within a fixed distance of the pith, and site quality for tolerant species can be represented by the slope of the height-age curve at a standard height.
Al was one of the original inhabitants of the Moscow Forestry Sciences Laboratory when it opened in 1963. In the early 1960's Al was instrumental in introducing point-sampling methods to the National Forest Survey. His interest in forest survey and forest growth and yield led to involvement in timber trends forecasting for the Forest Survey. He wrote the first FORTRAN version of the TRAS growth simulation program that was used in the 1962 assessment. Working with Bob Pope from the Pacific Northwest Station, he introduced a density feedback mechanism to slow diameter growth and increase mortality rates when canopies closed. TRAS was not designed to work with specific stand conditions. It projected stand tables in 2-inch diameter classes, and data were generally aggregated across broad cover classes prior to projection. These projection methods were not amenable to the diverse stands that characterize the Intermountain Northwest forests. Al credits the experience with TRAS as a major motivation for development of the Prognosis Model.
Although Al is quick to recognize other contributors to system development, the Prognosis Model has come to be synonymous with Al Stage. He introduced Prognosis in 1973 and spent most of the next 20 years engaged in a broad array of cooperative associations that gave substance to the system. While TRAS was the primary motivation for development of Prognosis, the design was undoubtedly colored by Al's experiences with grand fir site index and his studies of the relationship between white pine pole blight and climate. He recognized that site index was not an appropriate basis for growth projection in irregular stands and that forest projections used for planning must account for the impacts of forest pests. Rather than using site index, productivity was indexed to Daubenmire's land classification, which was based on climax vegetation. Other novel concepts introduced with Prognosis were the omission of stand and tree age from growth models, the use of relative size and crown ratio to represent tree vigor, and a calibration procedure that used the past growth experience to tune growth models for an individual stand.
Al's vision for the Prognosis Model included the ability to simulate growth and management for almost any forested site, with treatment options limited only by imagination. As a result, the system includes extensive thinning options that can be keyed to almost any stand or tree condition. Furthermore, predictions of shrub cover provide links to wildlife and water resource models, and a regeneration establishment model permits evaluation of regeneration treatments and exploration of forest succession. Prognosis was designed to work with specific stand inventories, and Al collaborated with Jack Alley at the USDA Forest Service Northern Region to develop a complementary management planning inventory system. Because each projection is based on an actual inventory, the initial conditions exactly match the inventory compilation. Because stands are projected individually before information is aggregated, the estimates can be tied to a specific piece of ground and estimates of variability in projected conditions are retained. The prediction approach recognizes that growth is a deviation amplifying process and that there is considerable unexplained variation in the growth estimators. If all trees are projected at the mean for the extant conditions, then stand attributes are generally underestimated. Thus residual variation is systematically applied to tree growth predictions to minimize bias in projected basal area and volume estimates. One of the tenets underlying Al's approach to research is that variation is the raw material of forest planning-projection and planning methods that ignore variability are less than optimal.
The Prognosis Model was also designed to incorporate effects of pests on stand development. Thus, when "big bug" programs funded large-scale research on western pine beetles, Douglas-fir tussock moth, and spruce budworm, Prognosis was linked to pest models to provide substrate for the pest populations. This association also allowed for accounting of pest impacts on projected yield estimates. Now there are additional Prognosis extensions that accommodate the effects of dwarf mistletoes, root diseases, white pine blister rust, and fire. Another innovation was a parallel processing extension that allows for simultaneous projection of a collection of stands where treatment can be scheduled in accordance with broad policy rules. The parallel processor permits consideration of adjacency constraints on management actions and allows contagion effects in fire and pest population simulations. Prognosis, the many pest extensions, and the parallel processor were amazing developments, particularly when considered against the backdrop of early 1970's computing facilities, where analysis was expensive and turn around on a single computer job could be 24 hours or more.
In the 1990's, political winds fanned the evolution of Prognosis into the Forest Vegetation Simulator, and variants were developed to represent most of the forest conditions in the U.S. Al did not create Prognosis single-handedly - there have been hundreds of people involved in the data collection and analysis of component models and the coding of the myriad of variants and extensions. But, Al's vision, his quiet but persuasive prodding and his firm grasp of biophysical, mathematical, and statistical concepts have served as a strong foundation for the system. Along the way, Al has achieved many awards, including the Distinguished Alumni Award from the University of Michigan School of Natural Resources, the USDA Superior Service Award, the Forest Service Forest Insect and Disease Award for Research Excellence, and Fellow of the Society of American Foresters. The extent of his influence is exemplified by the more than 50 scientists with whom he has co-authored papers. The breadth of his contribution is reflected in the array of topics addressed in those papers, including basic mensuration, sampling, hydrology, plant moisture stress, climate and disease, insect population dynamics, root disease, fertilization, fire effects, and a host of papers on techniques for modeling stand and tree dynamics. The extent of Al's influence on the direction of current forestry research will never be fully appreciated. I personally stand in awe of Al's achievements. I am still exhausted by the fervor and pace with which he attacks forestry research. But I am immensely pleased to have had the opportunity to play a small part in his dream.
William R. Wykoff
USDA Forest Service--RMRS
wwykoff@fs.fed.us
