Population-averaged and subject-specific approaches for clustered categorical data

Oliver Schabenberger
Stat Collaborat Inc,1710 Rhode Isl Ave Nw,Suite 200,Washington,DC 20036

Timothy G. Gregoire

Virginia Polytech Inst & State Univ,Coll Forestry & Wildlife Resources,Blacksburg,VA 24061

Abstract
Modeling clustered categorical data based on extensions of generalized linear model theory has received much attention in recent years. The rapidly increasing number of approaches suitable for categorical data in which clusters are uncorrelated but correlations exist within a cluster, has caused uncertainty among applied scientists as to their respective merits and demerits. Upon centering estimation around solving an unbiased estimating function for mean parameters and estimation of covariance parameters describing within-cluster or among-cluster heterogeneity, many approaches can easily be related. This contribution describes a series of algorithms and their implementation in detail, based on a classification of inferential procedures for clustered data. Schabenberger, Gregoire 1996. J. Statist. Comput. Simulation 54 (1-3): 231-253.

Author Keywords
generalized mixed linear model, Pseudo-likelihood, estimating equations, nomenclature

Key Words Plus
Generalized linear-models, longitudinal data-analysis, mixed effects models, multivariate methods, binary data, regression, parameters, outcomes

A versatile growth model with statistically stable parameters

Jon Schnute

Dept. Fisheries and Oceans, Resources services Branch, Pacific Biological Station, Nanaimo, B.C. V9R 5K6. Canada

Abstract

This paper presents a new comprehensive growth model which includes numerous historical models as special cases. The new model is derived from a concise biological principle which, unlike earlier theories, relates to growth acceleration. Properties of growth curves, such as asymptotic limits or inflection points, are shown to be incidental in this new context. Possible submodels include not only asymptotic growth (such as Von Bertalanffy, Richards, Gompertz, or Logistic growth) but also linear, quadratic, or exponential growth. By simple analysis of variance, the observed data can be used directly n deciding which type of model is most appropriate. The new model is cast in terms of parameters which have stable statistical estimates. From this perspective, it is shown how earlier formulations sometimes result in an endless computer search for optimal parameter estimates. Schnute 1981. Can. J. Fish Aquat. Sci. 38:1128-1140.

Key words

Growth, asymptotic growth, von Bertalanffy, Richards, Gompertz, Logistic, length at age, weight at age, nonlinear parameter estimation

Biomass estimation for temperate broadleaf forests of the United States using inventory data

Paul Schroeder

Dynamac Corp,US EPA,Natl Hlth & Environm Effects Res Lab,Western Ecol Div,200 SW 35th St,Corvallis,OR 97333

Sandra Brown

US EPA,Natl Hlth & Environm Effects Res Lab,Western Ecol Div,Dynam Corp,200 SW 35th St,Corvallis,OR 97333

Mo Jiangming

Dinghushan Arboretum,Guangdong 526070,Peoples R China

Richard Birdsey

US Forest Serv,NE Forest Expt Stn,Radnor,PA 19087

Chris Cieszewski

D.B. Warnell School of Forest Resources, University of Georgia ,Athens, GA 30602, USA

Abstract
A potentially valuable data source for estimating forest biomass is forest volume inventory data that are widely collected and available throughout the world. In this paper we present a general methodology for using such data to reliably estimate aboveground biomass density (AGED) and to develop expansion factors for converting volume directly to AGED from USDA Forest Service Forest Inventory and Analysis (FIA) data. Growing stock volume inventory data and stand tables were combined with ndependently developed biomass regression equations to estimate AGED and to calculate biomass expansion factors (BEF: factors that convert volume to mass, accounting for noncommercial components) for the extensive oak-hickory and maple-beech-birch forest types of the eastern United States. Estimated aboveground biomass for both forest types ranged between 28 and 200 Mg ha(-1). Expansion factors decreased from more than 4.0 at low growing stock volume to nearly 1.0 when growing stock volume was as high as 190 m(3)ha(-1), consistent with theoretical expectations. In stands with low AGED (< 50 Mg ha(-1)), small diameter trees (< 10 cm diameter) contained up to 75% of the AGED in trees greater than or equal to 10 cm diameter; this proportion dropped to < 10% for stands with AGED > 175 Mg ha(-1). The similarity of our results for two major forest types suggests that they may be generally applicable for estimating AGBD from inventory data for other temperate broad leaf forests. Further, the pattern between BEF and stand volume was similar to that obtained for tropical broadleaf forests, except that tropical forests generally had larger BEFs than temperate forests at a given volume. The implications of these results suggest that a recent assessment of forest biomass in developed countries is too low. Schroeder, Brown et al. 1997. For. Sci. 42 (3): 424-434.

Author Keywords
forest biomass, expansion factors, forest inventory, global carbon cycle

Key Words Plus
Carbon Budget , Tropical forests

Investigating parameters of growth equations

Shvets V

Univ Arkansas, Sch Forest Resources, Arkansas Agr Expt Stn,Monticello,AR 71656

Boris Zeid

Univ Arkansas, Sch Forest Resources,Arkansas Agr Expt Stn, Monticello,AR 71656

Abstract
Two differential forms of growth equations, called the power decline, or PD form, and the exponential decline, or ED form, generate classic growth equations (such as the logistic, Chapman-Richards, Korf) and many other integral forms. Having a full range of these integral solutions allows us to classify them, establish requirements to their parameters, and relate these parameters and initial values (starting age and tree size). Comparisons with data confirm theoretical results. Some applications of the results are discussed. Shvets, Zeide 1996. Can. J. For. Res. 26 (11): 1980-1990 Nov 1996.

Potential impacts of increased management intensities on planted pine growth and yield and timber supply modeling in the South

Jacek P. Siry

N Carolina State Univ, Dept Forestry, Box 8008, Raleigh, NC 27695 USA

Frederick W. Cubbage

N Carolina State Univ, Dept Forestry, Raleigh, NC 27695 USA

Malmquist AJ

Forest Technol Grp, Informat Technol, Summerville, SC 29483 USA

Abstract
Five management intensity classes were developed and used to estimate the potential planted pine growth and yield levels to be applied in the 2000 Rangeland Renewable Resource Act (RPA) assessment. These planted pine yields are compared to empirical pine yields, developed as a part of the Subregional Timber Supply model inputs and based on Forest Inventory Analysis (FIA) data. These comparisons indicate that planted pine yields are much greater than empirical FIA data, e.g., up to 94 percent greater than current empirical yields (2,361 ft.(3)/ac.) at the highest management intensity (4,587 ft.(3)/ac.) at age 25. Planted pine yields are also substantially higher than those used in the 1993 RPA assessment. If realized, such productivity increases could prevent potential timber shortages in the South. Financial analyses indicate that intensified forest management offers attractive returns. For medium-quality sites, the calculated net present values are only $290/ac. for the basic empirical southern pine growth rates. They are $359/ac. for traditional planted pine management, $526/ac. with improved genetics, $581/ac. with fertilizers, $798/ac. with fertilizers and herbicides, and $939/ac, with repeated applications of fertilizers and herbicides. Siry, Cubbage 2001. Forest Prod. J. 51 (3): 42-48. 51 (3): 42-48.

Asymptotic site-index curves, fact or artifact

Victor G. Smith

Univ Toronto, Fac Forestry,Toronto M5S 1A1,Ontario,Canada

Abstract

In the absence of permanent sample plot data, site index curves and equations showing the height/age relationship for a tree species growing on a given site have been derived suing stem analysis and temporary sample plot data. Initially the guide-curve method using temporary sample plot data (Bruce and Schumacher 1950) was employed to produce anamorphic site-index curves. More recently mensurationists (e.g. Heger 1988, and Payandeh 1978) have used stem analysis data dn various regression techniques to produce polymorphic curves. This study suggests that the asymptotic site-index curves (i.e. curves that tend to level off within the life of the stand) that have been produced for black spruce may be the result of using mensurational and sampling techniques in forests where site-index is correlated with age. Also the regression techniques commonly used to estimate the parameter in site-index equations do not allow for error in both the dependent variable (height) and an independent variable (site). As a consequence the site-index curves produced by these techniques are asymptotic whereas height estimates obtained from 40 black spruce permanent sample plots do not exhibit any strong asymptotic height properties for stands up to 180 years of age. Smith 1984 .Can. J. For. Res. 17:1181-1189.

Spatial variation and patterns of soil microbial community structure in a mixed spruce-birch stand

Peter Saetre

Department of Ecology and Environmental Research, Swedish University of Agricultural Sciences, Box 7072, Uppsala S-750 07, Sweden

Erland Baath

Department of Microbial Ecology, Ecology Building, Lund University, Lund S-223 62, Sweden

Abstract
To explore the spatial variation of the soil microbial community within a mixed Norway spruce-birch stand, and to test if the spatial patterns of the microbial community are related to the position of trees, we sampled the forest floor at two spatial scales and used the phospholipid fatty acid (PLFA) patterns as indicators of the microbial community structure. Of the 32 most common PLFAs, 20 (62%) were clearly spatially autocorrelated, and the limit of spatial dependence (range) varied between 1 m and 11 m. The variation in the community structure was examined by subjecting the PLFAs to a principal component analysis. The first two principal components described variation structured at two different spatial scales. The range of the microbial community for the first component was 4.6 m, whereas for the second component it was only 1.5 m. The microbial community was influenced by the position of the trees. Spruce trees had a much stronger influence on PLFA patterns than birch trees, and the first principal component, as well as 12 PLFAs, was influenced by spruce trees. Several branched PLFAs, characteristic of Gram-positive bacteria, loaded negatively on the second principal component. These PLFAs represent a complex of associated microorganisms that aggregated in small patches away from birch trees. A comparison with a laboratory experiment suggests that although the tree species differ in their influence on soil moisture and ground vegetation, their influence on the microbial community were, to a large extent, connected to the quality of soil organic matter associated with the two trees. Saetre and Baath 2000. Soil Biol. Biochem. 32: 909-917.

Author Keywords
spatial variation, soil microbial community, phospholipid fatty acids, tree influence, Picea abies, Betula pubescens

KeyWords Plus:
fatty-acid analysis, forest soils, nitrogen mineralization, biomass, heterogeneity, vegetation, ph, variability, bacterial, chemistry

Spatial variation of nitrate-N and related soil properties at the plot-scale

Stenger R

Institute of Soil Ecology, GSF-National Research Center for Environment and Health, D-85764 Neuherberg, Germany

Priesack E

Lincoln Environmental, Private Bag 3062, Hamilton, New Zealand

Beese F

Institute of Soil Science and Forest Nutrition, University of Gottingen, Germany

Abstract:
Neglecting the spatial variation in soil nutrient status may result in unused yield potential and in environmental damage. Site-specific management has been suggested to reduce inappropriate fertilization that can adversely affect soil, ground and surface water. Decision criteria for determining variable-rate nitrogen fertilization are, however, lacking. This paper analyses the spatial variation of nitrate nitrogen (NO3-N) and soil properties related to the N cycle at the plot-scale, Three 50 X 50 in plots were sampled in nested sampling designs of varying complexities. Classical statistics revealed a characteristic ranking in the variability of soil properties. Geostatistical analysis of the NO3-N data from two plots showed that the small-scale variation found in one small subgrid was not typical for the small-scale variation in the entire plot, indicating bias in the sampling design. A trend component was found in the NO3-N data and, consequently, the minimal requirement for the regionalized variable theory was not fulfilled. Problems due to design were overcome with a more complex nested sampling at the third plot. However, the spherical model fitted to the NO3-N data of the first year explained only 21% of the total variance, whereas a pure nugget effect was observed in the second year. The water content data also showed a low structural variance, which was different in the two years, In contrast, two thirds of the variance of total carbon (C) and total nitrogen (N) could be explained by the fitted models. Seasonal variations, such as varying duration of snow cover, and extrinsic management effects, such as growing of a cover crop, may have contributed to the observed differences in variability between the years. Due to the low proportion of structural variance and the observation that spatial distribution was not stable with time, geostatistical analysis of NO3-N and water contents data added only little information to classical statistical analysis. However, geostatistical analysis of total C and N contents provided a useful means to calculate spatial distribution patterns of these properties. Stenger, Beese et al. 2002. Geoderma. 105: 259-275.

Author Keywords:
site-specific management, spatial variation, temporal variation, heterogeneity

KeyWords Plus:
temporal persistence, field-scale, variability, patterns, nitrogen, mineralization, availability, dependence, management

Using competitive stress index to estimate diameter growth for thinned Douglas-fir stands

S. H. Smith

Potlatch Corporation, Lewiston, Idaho

J. F. Bell

Department of Forest Management, Oregon State University, Corvallis, OR 9733

Abstract

A set of linear growth equations was developed using initial tree diameter (DBH), initial competitive stress index (CSI), and change in CSI due to thinning as independent variables to predict periodic tree-diameter growth response for a 3- and 4-year growth period. Data from a regional levels-of-growing-stock study in young-growth Douglas-fir [Pseudotsuga menziesii (Mirb.) Franco] provided the data base for computing the CSI values. Functions including all three independent variables accounted for the greatest variation in periodic diameter increment for both growth periods. Moreover, including transformations of initial CSI and change in CSI in the model provided significant improvements over a model based only on functions of initial DBH. Smith and Bell 1983. For. Sci. 29: 491-499.

Additional key words

Intertree competition, growing space, influence zone, stand models

Yield prediction for unthinned natural slash pine stands

H.T. Schreuder

Southeastern Forest Experiment Station, U.S. Forest Service, Research Triangle Park, North Carolina

W.L. Hafley

School of Forest Resources, North Carolina State University, Raleigh, North Carolina

F.A. Bennett

Abstract

Equations are presented for predicting number of trees, mean height, and total volume outside bark by diameter class when stand age (A), site index (SI), and current number of trees per hectare (N) are known in unthinned natural stands of slash pine. The Weibull distribution was fitted to diameter frequency data, then suitable height-diameter and volume-D²H models were developed to calculate the Weibull parameters for the distributions of height and volume. Parameters for diameter and height distributions can be predicted form A, SI, and N and these equations are incorporated with the volume- D²H model in a computer program that generates a diameter frequency distribution, mean height, and total volume per acre or hectare by diameter class. Schreuder, Hafley et al. 1979. For. Sci. 25: 25-30.

Additional key words

Weibull distribution, functional models, diameter distribution, height distribution, volume distribution

Yield of unthinned red pine plantations at the Petawawa forest experiment station

W.M. Stiell

Forest Management Institute Ottawa, Ontario

A.B. Berry

Petawawa Forest Experiment Station Chalk River, Ontario

Abstract

Yield tables for high-survival, unthinned plantations of red pine were prepared to age 50 years from planting, by 5-year age classes, for eight planted spacing and five site index classes. Each table presents the numbers of trees and the basal area per acre, the mean d.b.h., and the total, merchantable cubic, and board foot volumes per acre. Site index curves and diameter distribution data are also presented. Stiell and Berry 1973. Dept. Environ. Can. For. Serv. Publ. NO. 1320. 17p.

Michigan's forests, 1993: an analysis

Thomas L. Schmidt, John S. Spencer, Jr.

Forest Inventory and Analysis Program at the North Central Forest Experiment Station, St. Paul, Minnesota.

Robin Bertsch

Forest Management Division of the Michigan Department of Natural Resources, Lansing, Michigan.

Abstract

The fifth inventory of Michigan's forests reports 36.4 million acres of land, of which 19.3 million acres are forested. This bulletin presents an in-depth analysis of the forest resources and contains detailed tables of area, volume, growth, removals, mortality, and ownership. Schmidt and Spencer, Jr. 1997. USDA For. Serv. Res. Rep. NC-179. 101p.

Key Words

forest area, timber volume, growth, removals, mortality

Determination of spatial continuity of soil lead levels in an urban residential neighborhood

Nita J. Shinn, John Bing-Canar, Michael Cailas

School of Public Health, University of Illinois at Chicago, Chicago, Illinois

Nicholas Peneff

School of Public Health, University of Illinois at Chicago, Chicago, Illinois; - Public Health & Safety, Inc., Chicago, Illinois

Helen J. Binns

Children’s Memorial Hospital, Northwestern University, Chicago, Illinois

Abstract
This study uses geostatistical techniques to model and estimate soil lead levels in an urban, residential neighborhood. Sixty-two composite soil samples (median 1773 ppm; range 175 to 7953 ppm) in a four-block area of brick and stone homes were obtained. The spatial continuity of soil lead levels was modeled with a semi-variogram, which was then used to estimate lead levels at unsampled locations, a process called kriging. Because soil lead levels were spatially correlated, it is likely that a "nonrandom" process generated the lead distribution found. This finding signifies the existence of lead sources which were tentatively identified on historical maps of the area and from past traffic volume patterns. The distribution of kriged estimates of soil lead levels provides an explanatory tool for exploring and identifying potential sources and mag be useful for targeting urban soil abatement efforts. Shinn, Bing-Canar et al. 2000. Environmental Research. 82: 46-52.

Author Keywords
soil, lead, mapping, geostatistics, urban

KeyWords Plus
blood lead, children, smelter, exposure, plants

Adapting an individual tree growth model for Norway spruce (Picea abies L. Karst.) in pure and mixed species stands

Hubetr Sterba, Astrid Blab

Institute of Forest Growth Research, University of Agricultural Sciences, Peter Jordanstrabe 82, A-1190 Wien, Vienna, Austria

Klaus Katzensteiner

Institute of Forest Ecology, University of Agricultural Sciences, Peter Joudanstrabe 82,A-1190, Wien, Vienna, Austria

Abstract
PROGNAUS is an individual tree growth model, containing a major sub-model which depicts basal area growth as a function of tree size, competition and a set of diagnostic site variables. The coefficients of this model have previously been estimated using the data from the Austrian National Forest (ANF) inventory. The examination of 60 new sample plots in two Austrian growth districts showed that the increment of Norway spruce (Picea abies L. Karst.) in pure spruce stands was overestimated by this model while it was underestimated in mixed species stands with European beech (Fagus sylvatica L.).

In order to adapt PROGNAUS to account for this effect, the data of the 60 plots were used to modify PROGNAUS by adding quantitative soil characteristics and the percentage of Not-way spruce as variables. The markedly improved model accounts for the effect that in a given stand, below a certain competition level, Norway spruce grows better in the pure stand than in an otherwise comparable mixed species stand. These small spruce trees in mixed stands have a shallow rooting system and therefore suffer from water stress, which is caused by the competing European beech with its higher water consumption. and deeper rooting system.

Adding the Norway spruce proportion as an independent variable into the original model and re-parameterising it with the data of the ANF inventory, reveals that this effect can be generalized to some extent, but differs in magnitude depending on the geological substrate. Sterb, Blab et al. 2002. For. Ecol. Manage. 159: 101-110.

Author Keywords
individual tree growth model, mixed species stands, Picea abies L. Karst., Fagus sylvatica L.

KeyWords Plus
forest stands, beech

Forest floor evaporation in a dense Douglas fir stand

Schaap MG, Bouten W

Landscape and Environmental Research Group, University of Amsterdam, Nieuwe Prinsengracht 130, 1018 VZ Amsterdam, The Netherlands

Abstract:
Forest floor evaporation was measured with an accurate weighing lysimeter during 44 days in early spring and summer. The Penman-Monteith approach was used to model the evaporation rates as well as the temperature difference between forest floor surface and air at Im height. Values of resistance parameters were slightly different when the Penman-Monteith model was optimized for measured evaporation rates or for measured temperature differences. These discrepancies were partly due to field variability in forest floor water contents but also because our approach considered the forest floor to be isothermal. With the appropriate parameter sets, the model was able to predict measured hourly forest floor evaporation rates and surface temperature dynamics satisfactorily.

We show that in the forest discussed in this paper the Penman-Monteith ventilation term dominates over the available energy term. As a result the evaporation flux is matched by an almost equal, sensible heat flux but in opposite direction. Forest floor water content dynamics have a strong control over the evaporation flux. Spatial variability in forest floor water contents cause the 44-day average forest floor evaporation to range from 0.19 mm d(-1) in a dry part of the forest to 0.3 mm d(-1) in a wet part with 0.23 mm d(-1) as a site representative value. Schaap and Bouten 1997. J. Hydrology. 193: 97-113.

KeyWords Plus:
canopy, evapotranspiration, transpiration, model

Satellite-derived vegetation index and cover type maps for estimating carbon dioxide flux for arctic tundra regions

Douglas Stow, Allen Hope, William Boynton, Stuart Phinn

Department of Geography, San Diego State University, San Deigo, CA 92182-4493, USA

Donald Walker, Nancy Auerbach

Institute of Arctic and Alpine Research, University of Colorado, Boulder, CO 80309,USA

Abstract
The spatial variability and co-variability of two different types of remote sensing derivatives that portray vegetation and geomorphic patterns are analyzed in the context of estimating regional-scale CO2 flux from land surfaces in the arctic tundra, For a study area encompassing the Kuparuk River watershed of the North Slope of Alaska, we compare satellite-derived maps of the normalized difference vegetation index (NDVI) generated at two different spatial resolutions to a map of vegetation types derived by image classification of data from the Landsat multispectral scanner (MSS). Mean values of NDVI for each cover type stratum are unique (with the exception of moist acidic tundra and shrubland types), Based on analysis of semi-variograms generated for SPOT-NDVI data, most of the vegetation cover and landform features of this arctic tundra landscape have spatial dimensions of less than 1 km. Thaw lakes on the coastal plain and glacial depositional landforms, such as moraines in the foothills, seem to be the largest features, with vegetation units having dimensions no larger than 700 m. Frequency distributions of NDVI and vegetation types extracted for sampling transects flown by an aircraft sensing CO2 flux, relative to distributions for the entire Kuparuk River watershed, suggest a slight sampling bias towards greater cover of mesic wet sedge tundra and thaw lakes and associated lower NDVI values. The regional pattern of NDVI for the North Slope of Alaska corresponds primarily to differences between the two major physiographic provinces of this region. Stow, Hope et al. 1998. Geomorphology. 21: 313-327.

Author Keywords:
arctic tundra, remote sensing, carbon dioxide, regional scaling, satellite imagery, spectral vegetation indices

KeyWords Plus:
reflectance, ecosystems, co2, exchange

Modeling canopy structure and heterogeneity across scales: From crowns to canopy

Bo Song, Jiquan Chen, Paul V. Desanker, David D. Reed

Schoool of Forestry and Wood Products, Michigan Technological University, Houghton, MI 49931, USA

Gay A. Bradshaw

Forestry Science Laboratory, USDA Forestry, USDA Forest Service, Corvallis, Or 97331, USA

Jerry F. Franklin
College of Forest Resources, University of Washington, Seattle, WA 98195, USA

Abstract
Canopy studies have been limited in ecological investigations due to problems of canopy accessibility, and the lack of efficient sampling and modeling methods. The primary objective of this study was to develop an efficient modeling approach to describe the 3-dimensional, hierarchical structure of individual crown shells within stands and corresponding canopy patches. Crown shells were modeled based on crown ratio, maximum cardinal radius, vertical position, and shape. Canopies were represented by adding unique crowns to simulated point patterns of trees of known aggregation as measured by Pielou's index of nonrandomness. Canopy patches were delineated at multiple horizontal and vertical scales using the ARC/INFO geographic information system (GIS). The patterns of canopy patches are clearly variable and scale dependent. Canopy patterns become more diverse at broader horizontal scales, and change greatly from the lower to the upper canopies. The modeling approach used in this study has general utility in characterizing 3-dimensional canopies of many types of forests. Song, Chen et al. 1997. For. Ecol. Manage. 96: 217-229.

Author Keywords
crown, canopy, simulation, GIS, scale, spatial pattern

Can ^239+240Pu replace ¹³⁷Cs as an erosion tracer in agricultural landscapes contaminated with Chernobyl fallout?

Schimmack W

GSF-National Research Center of Environment and Health, Institute for Radiation Protection, P.O. Box

1129, D-85758 Neuherberg, Germany

Auerswald K

Department of Soil Science, Technical University MuÈnchen/Weihenstephan, D-85350 Freising, Germany

Bunzl K

GSF-National Research Center of Environment and Health, Institute for Radiation Protection, P.O. Box

1129, D-85758 Neuherberg, Germany

Abstract
Erosion studies often use Cs-137 from the global fallout (main period: 1953-1964) as a tracer in the soil. In many European countries, where Cs-137 was deposited in considerable amounts also by the Chernobyl fallout in 1986, the global fallout fraction (GF-Cs) has to be separated from the Chernobyl fraction by means of the isotope Cs-134. In a few years, this will no longer be possible due to the short half-life of Cs-134 (2 yr). Because GF-Cs in the soil can then no longer be determined, the potential of using Pu239+240 as a tracer is evaluated. This radio nuclide originates in most European countries essentially only from the global fallout. The activities and spatial distributions of Pu and GF-Cs were compared in the soil of a steep field (inclination about 20%, area ca. 3 ha, main soil type Dystric Eutrochrept), sampled at 48 nodes of a 25x25 m(2) grid. The reference values were determined at 12 points adjacent to the field. Their validity was assured by an inventory study of radiocaesium in a 70 ha area surrounding the field sampling 275 nodes of a 50x50 m(2) grid. In the field studied, the activity concentrations of GF-Cs and Pu in the Ap horizon were not correlated (Spearman correlation coefficient R = 0.20, p > 0.05), and the activity balance of Pu differed from that of GF-Cs. Whereas no net loss of GF-Cs from the field was observed as compared to the reference site, Pu was more mobile with an average loss of ca. 11% per unit area. In addition, the spatial pattern of GF-Cs and Pu in the field differed significantly. The reason may be that due to their different associations with soil constituents, Pu and Cs represent different fractions of the soil, exhibiting different properties with respect to erosion/deposition processes. This indicates that both radionuclides or one of them may not be appropriate to quantify past erosion. When tracer losses are used to calibrate or verify erosion prediction models, systematic deviations may not only stem from model shortcomings but also from tracer technique. Schimmack, W., K. Auerswald et al. 2001. J. Environ. Radioactiv. 53: 41-57.

Author Keywords
erosion, plutonium, radiocaesium, Chernobyl fallout, global fallout

KeyWords Plus
spatial variability, sequential extraction, grassland soil, Cesium-137, Am-241, association, Pu-239+240, radiocesium, limitations, plutonium

Process-based models and forest management

Risto Sievanen

Department of Mathematics, The Finnish Forest Research Institute Unioninkatu 40 A, 00170 Helsinki Finland

Thomas E. Burk

Department of Forest Resources, University of Minnesota 1530 N Cleveland Avenue, St. Paul, MN 55108 USA

Abstract

The application of process-based (PB) growth models in forest management has been quite limited so far. However, In principle these models have the potential for being highly applicable in solving various forest management problems. Before this potential can be realized, the models have to meet the special requirements arising form the intended application. In this paper the capabilities of PB models to meet these requirements is studied. Sievanen and Burk. Dept. Math. Fin. For. Res. Ins. & Dept. For. Res., Univ. Minn. 13p.

Parameter estimation in a photosnthesis-based growth model

Risto Sievanen

Department of Mathematics, The Finnish Forest Research Institute Unioninkatu 40 A, 00170 Helsinki Finland

Thomas E. Burk, Alan R. EK

Department of Forest Resources, University of Minnesta 1530 N Cleveland Avenue, St. Paul, MN 55108 USA

Abstract

A growth model for an individual tree in a forest stand based on photosynthesis and respiration relationships is briefly described. Individual tree growth data (six measurements of diameter) of red pine (Pinus Resinosa Ait.) stands are used to estimate a subset of the parameters in the model using a globally, simultaneous approach. The loss function is composed of prediction errors for individual trees at the times of measurement and is a nonlinear function of the parameters. Estimates for parameters are compared with values found in the literature. Use of alternative loss function definitions provides insight into model component stability. Sievanen and Burk. P.345-352. IUFRO.

Construction of a stand growth model utilizing photosynthesis and respiration relationships in individual trees

Risto Sievanen

Department of Mathematics, The Finnish Forest Research Institute Unioninkatu 40 A, 00170 Helsinki Finland

Thomas E. Burk, Alan R. EK

Department of Forest Resources, University of Minnesta 1530 N Cleveland Avenue, St. Paul, MN 55108 USA

Abstract

A discrete growth model for an individual tree in a forest stand is presented. The model is based on simplified photosynthesis and respiration relationships. Implementation of the model also requires functions for change in live crown ratio and stems per unit area. Tree growth predictions are expressed in terms of biomass and basal area. A preliminary evaluation of the model is made using data from the literature on the growth of even-aged stands of red pine (Pinus resinosa Ait.). The model produced both short- and long-term trends consistent with stand growth data. The evaluation suggests further possibilities for model component refinement. Sievanen, Burk et al.1988. Can. J. For. Res. 18: 1027-1035. 1988.

Comparison of diameter growth and crown competition factor in loblolly pine plantations

Mike R. Strub

School of Forest Resources, Pennsylvania State University, University Park, PA 16802

Richard B. Vasey, Harold E. Burkhart

Division of Forestry and Wildlife Resources, Virginia Polytechnic Institute and State University (VPI & SU) Blacksburg, VA 24061

Abstract

Growth in diameter at breast height (dbh) in loblolly pine plantations was compared with that of open-grown trees. Data were collected in 240 plantations and from 81 open-grown pines. The age at which average diameter growth in plantations diverges from that of open-grown trees was estimated for different site index and stand density combinations and compared to the age at which the crown competition factor (CCF) is predicted to reach 100. Average diameter growth in stands consistently differed from open-grown tree diameter growth and year after the CCF reaches 100. This result is potentially useful for forest growth simulation. Strub, Vasey et al. 1975. For. Sci. 21: 427-431. 1975.

Additional key words

Pinus taeda, competition.

Volume estimation by diameter class using bootstrap variance analysis

Hans T. Schreuder, G.E. Brink

Rocky Mountain Forest and Range Experiment Station, USDA Forest Service, 240 West Prospect, Fort Collins, CO 80526 (U.S.A.)

Abstract

Bootstrap estimation is used to provide standard errors of estimates for volume by diameter class in a test data set of the IV^th Dutch Forest Survey. Such standard errors of estimates cannot currently be obtained by classical methods. Schreder and Brink 1987. For. Ecol. Manage. 20: 189-194. 1987

The jackknife - a useful statistical tool

Hans T. Schreuder, G.E. Brink

Rocky Mountain Forest and Range Experiment Station, USDA Forest Service, 240 West Prospect, Fort Collins, CO 80526 (U.S.A.)

Abstract

The Jackknife technique is a statistical approach for distribution-free bias reduction and standard error estimation. The most popular version of the technique is described and illustrated, and potential areas of application are indicated. The technique is primarily useful to estimate standard errors for complex sampling strategies where no good standard error estimates can be computed, and to decrease estimation bias in situations where for the complex sampling strategies now being developed by the forest survey units for updating timber estimates. Schreuder and Brink 1992. Timber MGT. 531-535. 1992

Alternative estimators for Point-Poisson sampling

Hans T. Schreuder, G.E. Brink

Rocky Mountain Forest and Range Experiment Station, USDA Forest Service, 240 West Prospect, Fort Collins, CO 80526 (U.S.A.)

R. L. Wilson

Regional Timber Office, Container Corporation of America, P.O. Box T, Fernanaidan Beach, FL 32034

Abstract

A new Jackknife-based variance estimate is derived for point-Poisson (point-3P) sampling. This variance estimate is shown to be superior to most estimates derived assuming point-pps with or without replacement sampling, for a population with a very strong linear relationship between the variable of interest and the covariate used for sample selection and estimation. Only a variance estimate suggested by L. Grosenbaugh is equally good. This variance estimate is recommended since it is simplest to computed. Both the standard point- Poisson and a proposed unbiased Jackknife estimate of total volume are about equally efficient. Schreuder, Brink et al. 1984. For. Sci. 30: 803-812. 1984

Additional key words

Jackknife, Raj estimator, negative variance estimates, pps with and without replacement, 3p

Forest tree production centres in Canada 1983

Smyth, J.H., A.J. Brownwright

Great lakes Forest Research Centre, Canadian Forestry Service, Department of The Environment

Abstract

A total of 103 public sector, private and forest industry nurseries produced planting stock for reforestation purposes in 1983, Of these, 47 grew bare-root stock and 83 grew containerized seedlings. Planned production of bare-root and containerized stock for 1983 was 281 million and 260 million seedlings, respectively. Since 1979, bare-root production expanded by 40%, or 81 million seedlings, and containerized stock increased by 140%, or 152 million seedlings. Total production is forecasted to increased to 285 million bare-root and 271 million containerized seedlings in 1984. Smyth and Brownwright 1984. Can. For. Serv. Dept. Env.,Inf. Rep. O-X-357. 54p. 1983.

Structural stage in Pacific Northwest forests estimated using simple mixing models of multispectral images

Donald E. Sabol Jr., Alan R. Gillespie, John B. Adams, Milton O. Smith

Department of Earth and Space Sciences, Campus Box 351310, University of Washington, Seattle, WA 98195, USA

Compton J. Tucker

Laboratory for Terrestrial Physics, NASA Goddard Space Flight Center, Greenbelt, MD 20771, USA

Abstract
We identified stages of regrowth in replanted clearcuts in D ouglas-fir/western hemlock forests in the Gifford Pinchot National Forest, southern Washington, USA, using a simple four-endmember constrained linear spectral mixing model applied to a multispectral Landsat Thematic Mapper image in order to separate and quantify spectral contributions from significant scene components. Spectral unmixing produces images of the fractional amount of the spectral endmembers, which were green vegetation, nonphotosynthetic vegetation, soil and "shade," which includes topographic shading and shadows. Changes in endmember fractions correspond to changes in surface composition (as viewed from above). Unresolved shadows comprise the primary indicator of canopy structure and hence, regrowth stage. To isolate shadows, shading predicted from a digital elevation model was removed from the image before mixture analysis. As stands regrow, the surface cover shifts from initial high proportion of slash and exposed soil, and low proportions of green vegetation and shadows, to low fractions of stems and soil with high fractions of green vegetation and shadows. This shift in surface composition defines a regrowth trend in an endmember fraction data space. Projection of data onto this line allows estimation of structural stage and stand age, and provides a framework for remote mapping and monitoring of forest regrowth. Field analysis of 495 forest stands, representing stand structural stages ranging from newly replanted cuts to stands greater than 250 years in age, was used to assess the accuracy and precision of predicted structural stages and stand ages. The spectral unmixing approach can be used to evaluate and monitor forest regrowth quickly over large areas of the Pacific Northwest forests, and is extendible to mapping basic vegetation community type as well as structural stage. Sabol Jr. and Gillespie et al. 2002. Remote Sens. Environ. 80: 1-16.

Key Words Plus
thematic mapper, tm data, canopy, reflectance, scattering, vegetation, variograms, age

Estimation of stand basal area growth and yield with a reverse logistic function

Riyaz A. Sadiq

Faculty of Forestry, University of Toronto, Toronto, Ont., Canada M5S 1A1

Abstract

Logistic curves have been used to study the growth of human and animal populations. Physical chemists have used it to study growth and senescence of chemical reactions. The present study investigates the application of the curve to forest stands estimating basal area growth and yield. Graphic analysis with the basal area growth data from permanent sample plots in nine (Pinus resinona Ait.) plantations of southern Ontario, exhibited reverse logistic trends. The parameters of the reverse of logistic function were estimated by nonlinear regression techniques. Freese's chi-square test was employed to determine the accuracy of the resulting estimates of basal area growth and yield. Results from the data used here indicate that the function not only fits the data well but also has high predictive power. Sadiq 1983. Can. J. For. Res. 13: 289-297.

Estimation of individual tree volumes with age and diameter

Riyaz A. Sadiq

Kashmir Forest Department Sheikh Bagh, Srinagar - l90001, Kashimir, India

Victor G. Smith

Faculty of Forestry University of Toronto, Toronto, Ont., Canada M5S 1A1

Abstract

Even though invention of high-precision equipment has reduced measurement errors associated with the estimation of heights of standing trees, height estimation is still an expensive and time-consuming operation. At limes it is difficult to determine especially in dense forests or in forests located in hilly terrain. The present study advocates a volume-age-diameter function to estimate volumes of individual trees. The technique presented here circumvents measurement of tree heights through the use of age which, however, restricts the application of the function to plantations or forests whose age is predetermined. Analyses with stem-analysis data from red pine (Pinus resinosa Ait.) plantations of southern Ontario indicate that the function estimates tree volumes more accurately than the standard methods commonly used. Sadiq and Smith. 1982. Can. J. For. Res. 13: 32-39.

Spatial variation and patterns of soil microbial community structure in a mixed spruce-birch stand

Peter Saetre

Department of Ecology and Environmental Research, Swedish University of Agricultural Sciences, Box 7072, Uppsala S-750 07, Sweden

Erland Baath

Department of Microbial Ecology, Ecology Building, Lund University, Lund S-223 62, Sweden

Author Keywords
spatial variation, soil microbial community, phospholipid fatty acids, tree influence, Picea abies, Betula pubescens

Key Words Plus
fatty-acid analysis, forest soils, nitrogen mineralization, biomass, heterogeneity, vegetation, ph, variability, bacterial, chemistry

Chromatin distribution pattern and cell functioning

T.S. Sahota, F.G. Peet, A. Ibaraki, S.H. Farris

Canadian Forestry Service, Pacific Forestry Center, 506 West Burnside Road, Victoria, B.C. Canada V8Z 1M5

Abstract

Chromatin distribution patterns of the nuclei of the follicular epithelial cell of Dendrocronus pseudotsugae Hopk. (Scolytidae: Coleoptera) were analyzed after creating their digital images. The five cell populations examined consisted of follicular epithelial cells at three different stages of structural and functional differentiation and two populations in which this differentiation was blocked. The results show that chromatin distribution patterns differ when cells are structurally and functionally different but are not different in cells that are structurally, developmentally, and functionally similar. The procedures are capable of detecting minute differences and appear useful for early detection of developmental and functional changes in relation to normal and pathological conditions. Sahota and Peet et al. 1985. Can. J. Zool. 64: 1908-1913.

Spatial variability of soil properties at different scales within three terraces of the Henares River (Spain)

A. Saldana, A. Stein, J.A. Zinck

ITC International Institute for Aerospace Survey and Earth Sciences, P.O. Box 6, 7500 AA Enschede, Netherlands

Abstract
This paper applies statistical and geostatistical procedures to a soil chronosequence on the terraces of the Henares River (NE Madrid) to analyse the spatial distribution of several soil properties and use the contribution of geostatistics to establishing a landscape evolution model of the area. Particle-size distribution, pH, calcium carbonate and organic carbon were analysed. Statistical procedures focus on analysing differences between terraces. Geostatistical procedures identify short- and medium-range variations within individual terraces at different scales. Standard transitive variogram models describe the properties of the younger terrace;whereas the linear intransitive model fits the majority of variograms of the older terrace. The analysis confirms and quantifies the decrease in variability of soil properties from young to old deposits, showing thus an increment of soil homogenisation with time. Ageing of the terraces causes the variables to show nontransitive variogram models with unbounded variances within the observation range. Saldana and Stein et al. 1998. Catena. 33: 139-153.

Author Keywords
chronosequence, spatial variability, variogram, soil homogenisation, spatial sampling

Key Words Plus
geostatistics, delineations, information, landscapes

On the applicability of Landsat TM images to Mediterranean forest inventories

Raymond Salvador

Centre de Recerca Ecologica i Aplicacions Forestals, Universitat Autonoma de Barcelona, 08193 Bellaterra, Barcelona, Spain

Xavier Pons

Departament de Geografia, Universitat Autonoma de Barcelona, 08193 Bellaterra, Barcelona, Spain

Abstract
Landsat TM images were used in combination with field measurements to create models for Mediterranean forests. Radiometric data from those images were related to field data from a forest inventory by means of regression analysis. Trials using plots with radiometrically homogeneous surroundings were carried out to evaluate the effect of high spatial heterogeneity frequently found in Mediterranean forests. Simple regression models were found to be consistent with the expected radiometric response of vegetation, and most of the multiple regression models fitted the observations sufficiently in order to make quantitative predictions for field variables from the remote sensing images. According to this study, however. these models should be regarded as exploratory models rather than fully operational ones. Spatially and non-spatially related factors are suggested as causes of the remaining dispersion of models created. Salvador, R., X. Pons. 1998. On the applicability of Landsat TM images to Mediterranean forest inventories. For. Ecol. Manage. 104: 193-208.

Author Keywords:
forest quantitative variables, radiometric response, predictive models, spatial heterogeneity

Key Words Plus
leaf-area index, thematic mapper, vegetation indexes, cover, spot, reflectance, stand, model, MSS

Soil carbon and soil physical properties response to incorporating mulched forest slash

Felipe G. Sanchez,

USDA Forest Service, Southern Research Station, P.O. Box 12254, Research Triangle Park, North Carolina 27709, United States

Emily A. Carter, John Klepac

USDA Forest Service, Southern Research Station, 520 Devall Drive, Auburn University, Alabama 36849-5418, United States

Abstract

A study was installed in the Lower Coastal Plain near Washington, North Carolina, to test the hypothesis that incorporating organic matter in the form of comminuted forest slash would increase soil carbon and nutrient pools and alter soil physical properties to favor pine growth. Two sites were selected, an organic and a mineral site, to compare the treatment effects on the different soil types. The mulching treatments included a surface broadcast mulch, a surface strip mulch, and a strip mulch and till. On the mineral site, the three treatments resulted in general decreases in soil bulk density, gravimetric soil water content, and soil strength. Soil carbon and nitrogen increased for all the treatments on the mineral site, with some significant differences between the treatments. The broadcast mulch and bed treatment resulted in an almost 100% increase in soil carbon and nitrogen. On the organic site, the treatments did not have a significant effect on soil physical properties or soil carbon and nitrogen. There was a consistent decrease in soil carbon and nitrogen on this site but these changes were not significantly different from those in the control treatment. Sanches and Carter et al. 2000. N. Z. J. For. Sci. 30: 150-168.

Keywords

mulch; till; soil carbon; bulk density; soil strength.

Forest floor evaporation in a dense Douglas fir stand

M.G. Schaap, W. Bouten

Landscape and Environmental Research Group, University of Amsterdam, Nieuwe Prinsengracht 130, 1018 VZ Amsterdam, The Netherlands

Abstract
Forest floor evaporation was measured with an accurate weighing lysimeter during 44 days in early spring and summer. The Penman-Monteith approach was used to model the evaporation rates as well as the temperature difference between forest floor surface and air at Im height. Values of resistance parameters were slightly different when the Penman-Monteith model was optimized for measured evaporation rates or for measured temperature differences. These discrepancies were partly due to field variability in forest floor water contents but also because our approach considered the forest floor to be isothermal. With the appropriate parameter sets, the model was able to predict measured hourly forest floor evaporation rates and surface temperature dynamics satisfactorily.

Key Words Plus
canopy, evapotranspiration, transpiration, model

The estimation of growth, site assessment and timber production forecasting from forest inventory in northern Ireland

R. Schaible

Forest Service, Department of Agriculture, Dundonald House, Upper Newtownards Road, Belfast BT4 3SB, Northern I r e land, UK

D.J. Kilpatrick

Biometrics Division, Department of Agriculture, Newforge Lane, Belfast BT9 5PX, Northern Ireland, UK

Abstract

There are 58,006 ha of state owned forest plantations in Northern Ireland managed as high forest. For inventory purposes the forest area is stratified into age classes and sampled by 2,100 randomly located permanent sample plots, some of which have existed from 1970. Measurements were made in the winter months between 1985 and 1988 to provide data for the estimation of standing volume and current annual increment, site assessment, timber production forecasting and the development o f yield models. Statistical methods were used to determine the optimum number off old and new plots to be measured in each stratum, and for the combination of successive measurements on individual plots to maximize the precision of the estimates.

Keywords

Forest inventory, stratified sampling, optimization

Contribution of increasing CO₂, and climate to carbon storage by ecosystems in the United States

David Schimel

Max-Planck-Institute for Biogeochemistry, Postfach 100164, D-07701 Jena, Germany

Jerry Melillo, Hanqin Tian

The Ecosystems Center, Marine Biological Laboratory, Woods Hole, MA 02543, USA

A. David McGuire

U.S. Geological Survey, Alaska cooperative Fish and Wildlife Research Unit, University of Alaska, Fairbanks, AK 99775-7020 USA

David Kicklighter

The Ecosystems Center, Marine Biological Laboratory, Woods Hole, MA 02543, USA

Timothy Kittel, Nan Rosenbloom

National Center for Atmospheric Research, P.O. Box 3000, Boulder, Co 80307-3000, USA

Steven Running, Peter Thornton

University of Montana, Missoula, MT 59812, USA

Dennis Ojima, William Parton, Robin Kelly

NREL, Colorado Sate University, Fort Collins, CO 80523-1499, USA

Martin Sykes

Plant Ecology, Lund University, Ekologihuset 223 62 Lund, Sweden

Ron Neilson

U.S. Department of Agriculture, Forest Service, Oregon State University, Forest Science Laboratory, 3200 Southwest Jefferson Way, Corvallis, OR 9733, USA

Brian Rizzo

Department of Enviornmental Sciences, University of Virginia, Charlottesville, VA 22903, USA

Abstract

The effects of increasing carbon dioxide (CO,) and climate on net carbon storage in terrestrial ecosystems of the conterminous United States for the period 1895-1993 were modeled with new, detailed historical climate information. For the period 1980-1993, results from an ensemble of three models agree within 25%, simulating a land carbon sink from CO, and climate effects of 0.08 gigaton of carbon per year. The best estimates of the total sink from inventory data are about three times larger, suggesting that processes such as regrowth on abandoned agricultural land or in forests harvested before 1980 have effects as large as or larger than the direct effects of CO, and climate. The modeled sink varies by about 100% from year to year as a result of climate variability. Schimel and Melillo et al. 2000. Science. 287: 2004-2006.

Can ^239+240Pu replace ¹³⁷Cs as an erosion tracer in agricultural landscapes contaminated with Chernobyl fallout?

Schimmack W

GSF-National Research Center of Environment and Health, Institute for Radiation Protection, P.O. Box

1129, D-85758 Neuherberg, Germany

Auerswald K

Department of Soil Science, Technical University MuÈnchen/Weihenstephan, D-85350 Freising, Germany

Bunzl K

GSF-National Research Center of Environment and Health, Institute for Radiation Protection, P.O. Box

1129, D-85758 Neuherberg, Germany

Abstract
Erosion studies often use Cs-137 from the global fallout (main period: 1953-1964) as a tracer in the soil. In many European countries, where Cs-137 was deposited in considerable amounts also by the Chernobyl fallout in 1986, the global fallout fraction (GF-Cs) has to be separated from the Chernobyl fraction by means of the isotope Cs-134. In a few years, this will no longer be possible due to the short half-life of Cs-134 (2 yr). Because GF-Cs in the soil can then no longer be determined, the potential of using Pu239+240 as a tracer is evaluated. This radionuclide originates in most European countries essentially only from the global fallout. The activities and spatial distributions of Pu and GF-Cs were compared in the soil of a steep field (inclination about 20%, area ca. 3 ha, main soil type Dystric Eutrochrept), sampled at 48 nodes of a 25x25 m(2) grid. The reference values were determined at 12 points adjacent to the field. Their validity was assured by an inventory study of radiocaesium in a 70 ha area surrounding the field sampling 275 nodes of a 50x50 m(2) grid. In the field studied, the activity concentrations of GF-Cs and Pu in the Ap horizon were not correlated (Spearman correlation coefficient R = 0.20, p > 0.05), and the activity balance of Pu differed from that of GF-Cs. Whereas no net loss of GF-Cs from the field was observed as compared to the reference site, Pu was more mobile with an average loss of ca. 11% per unit area. In addition, the spatial pattern of GF-Cs and Pu in the field differed significantly. The reason may be that due to their different associations with soil constituents, Pu and Cs represent different fractions of the soil, exhibiting different properties with respect to erosion/deposition processes. This indicates that both radionuclides or one of them may not be appropriate to quantify past erosion. When tracer losses are used to calibrate or verify erosion prediction models, systematic deviations may not only stem from model shortcomings but also from tracer technique. Schimmack, W., K. Auerswald et al. 2001. J. Environ. Radioactiv. 53: 41-57.

Author Keywords
erosion, plutonium, radiocaesium, Chernobyl fallout, global fallout

KeyWords Plus
spatial variability, sequential extraction, grassland soil, Cesium-137, Am-241, association, Pu-239+240, radiocesium, limitations, plutonium

Yields of four 40-year-old conifers and aspen in adjacent stands

Bryce E. Schlaegel

Southern Forest Experiment Station, Stoneville, Mississippi

Abstract

Volume and dry-weight yield comparison of 40-year-old planted red pine ( Pinus Resinosa Ait.), Jack pine (Pinus banksiana Lamb.), white spruce ( Picea glauca (Moench) Voss), black spruce (Picea Mariana (Mill.) BSP), and a naturally regenerated aspen (Populus tremuloides Michx) sucker stand show that red pine has the highest yields and black spruce the lowest. Volume yields of aspen, jack pine, and white Spruce are similar, but weight yields of white spruce are much less that wight yields of either aspen or jack pine. It is recommended that yield comparisons within and between species should be done on a weight as well as a volume basis because large differences in densities can occur. Schlaegel 1975. Can. J. For. Res. 5: 278-280.

Michigan's forests,1993: an analysis

Thomas L. Schmidt, John S. Spencer Jr.

Forest Inventory and Analysis Program at the North Central Forest Experiment Station, St. Paul, Minnesota

Robin Bertsch

Forest Management Division of the Michigan Department of Natural Resources, Lansing, Michigan

Abstract

Key words

Forest area, timber volume, growth, removals, mortality

A comparison of several artificial neural network classifiers for CT images of hardwood logs

Daniel L. Schmoldt

US Forest Service, Brooks Center, Virginia Tech, Blacksburg VA 24061-0503

Jing He

Communication Technology Group, COMSAT LAB, 22300 COMSAT Drive, Clarksburg MD 20871

A. Lynn Abbott

Bradley Department of Electrical Engineering, Virginia Tech, Blacksburg VA 24061-0111

Abstract

Knowledge of internal log defects, obtained by scanning, is critical to efficiency improvements for future hardwood sawmills. Nevertheless, before computed tomography (CT) scanning can be applied in industrial operations, we need to automatically interpret scan information so that it can provide the saw operator with the information necessary to make proper sawing decisions. Our current approach to automatically label features in CT images of hardwood logs classifies each pixel individually using a back-propagation artificial neural network (ANN) and feature vectors that include a small, local neighborhood of pixels and the distance of the target pixel to the center of the log. Initially, this ANN was able to classify clear wood, bark, decay, knots, and voids in CT images of two species of oak with 95% pixel-wise accuracy. Recently we have investigated other ANN classifiers, comparing 2-D versus 3-D neighborhoods and species-dependent (single species) versus species-independent (multiple species) classifiers using oak, yellow poplar, and cherry CT images. When considered individually, the resulting species-dependent classifiers yield similar levels of accuracy (96-98%). 3-D neighborhoods work better for multiple-species classifiers and 2-D is better for single-species. Under certain conditions there is no statistical difference in accuracy between single- and multiple-species classifiers, suggesting that a multiple-species classifier can be applied broadly with high accuracy. Schmoldt and He 1998. SPIE( the international Society for Optical Engineering). 3306: 34-43.

Keywords

industrial inspection, segmentation, computed tomography, image analysis, wood processing, lumber.

A Unified Approach to the Analysis of Fish Growth, Maturity, and Survivorship Data

Jon T. Schnute, Laura J. Richards

Department of Fisheries and Oceans, Biological Sciences Branch, Pacific Biological Station, Nanaimo, B.C. V9R 5K6

Abstract

This paper presents an approach to the analysis of growth, maturity, and survivorship data that unifies an extensive biometrics literature, particularly in fisheries and forestry. The model involves a probabilistic response y (0<y<1) to a positive factor x (0 <x < infinity). Because the relationship is asymptotic, the model can also be interpreted as a descriptor of biological growth, in which the asymptotic size is scaled to 1. The model generalizes earlier work by Gompertz, von Bertalanffy, Weibull, Richards, Chapman, and Schnute; it also extends the logit method introduced by Berkson. We identify reasonable curve families for the model, in which a transform parameter associated with x or y determines the curve shape, and we provide a complete analysis of curve types. We describe likelihood functions associated with two classes of data, typified by (1) size-at-maturity data and (2) time-mortality data. We illustrate the model’s application to both data classes and show that it provides a useful extension to past methods of analysis. Schnute and Richards. 1989. Can. J. Fish. Aqua. Sci. 47: 24-40.

Data uncertainty, model ambiguity, and model identification

Jon Schnute

Pacific Biological Station Nanaimo, B.C. V9R 5K6 CANADA

Abstract

This paper is based on a talk presented at the 1986 Resource Modeling Conference in Newport, Oregon. It addresses the problem of resource model identification, that is, the question: Which of various competing models correctly describes the resource? A major theme of the paper is the proposition that data uncertainty and model ambiguity are fundamentally intertwined. In some cases this relationship can be proved rigorously; consequently, model ambiguity can be inevitable. In other cases, promising identification techniques are suggested by modern statistical theory, and the paper examines some of these in the context of time series analysis. Throughout the paper, theoretical and practical examples illustrate both limitations and possibilities for using resource models as .valid descriptors of reality. Schnute 1987. Nat. Res. Modeling. 2: 159-212.

A general multivariate dose-response model

Jon Schnute, J.O.T. Jensen

Department of Fisheries and Oceans, Pacific Biological Station, Nanaimo, B.C. V9R 5K6

Abstract

This paper presents a general model for describing and predicting biological response to environmental factors, The theory embraces concepts and models used in earlier studies of the effects of toxicity on fish, and it extends these ideas in several directions. It can include an arbitrary number of factors, either detrimental or beneficial to the organism. Also, its statistical formulation allows (1 ) hypothesis testing to determine the relative importance of various factors and (2) calculation of critical factor levels that (with prescribed confidence) assure a community response no worse than a specified level, Our model is based on earlier work by Schnute and McKinnell, and, like theirs, its parameters have biologically meaningful interpretations. Our discussion includes an explanation of how the model relates specifically to toxicity studies. In a companion paper (Can. J. Fish. Aquat. Sci. 43: 1694-1 709), we use our model to investigate salmonid response to gas supersaturation and ancillary factors. Schnute and Jensen. 1986. Can. J. Fish. Aqua. Sci. 43: 1684-1693.

A biologically meaningful approach to response surface analysis

Jon Schnute, Skip McKinnell

Pacific Biological Station, Nanaimo, B.C. V9R 5K6

Abstract

Response surface analysis is commonly used to summarize the results of a complex biological experiment. The technique produces a quadratic function that relates experimental conditions (factors) to biological response, where the quadratic behavior allows for a possible optimum. Typically, the variables in this function are transformed to give the model greater flexibility. In fisheries literature, exponential transformations (with exponents presumed unknown) have been widely applied. Although properties of the quadratic model are well documented, the surface features that result from exponential transformations are not. Frequently, the practitioner obtains parameter estimates with widely varying magnitudes and contour plots that exhibit strange and confusing distortions. The parameters may bear no apparent relationship to the surface and, consequently, may appear to be useless pieces of information. This paper defines new parameters for the quadratic model with exponential transformations and demonstrates precisely the role of each parameter in determining the shape of the surface. The parameters have the added advantage that numerical methods to estimate them perform efficiently, thus avoiding convergence problems sometimes encountered in the past. Statistical error is discussed, analytically and intuitively in terms of model parameters. Evidence i s given that one of the transformations may be wrong in fisheries applications, and a reasonable alternative is proposed. Worked examples from historical literature illustrate all aspects of the new approach and show how both quantitative and qualitative errors can be avoided through biological understanding of the significance of every parameter. Schnute and McKinnell. 1984. Can. J. Fish. Aqua. Sci. 41: 936-953.

A new approach to estimating populations by the removal method

Jon Schnute

Pacific Biological Station Nanaimo, B.C. V9R 5K6 CANADA

Abstract

This paper presents a new approach to the use of removal data in estimating the size of a population of fish or other animals. The theory admits a variety of assumptions on how catchability varies among fishings including the assumption of constant catchability, which underlies most previous work. The methods here hinge on maximum likelihood estimation. and they can be used both to decide objectively if the data justify rejecting constant catchability and to determine confidence intervals for the parameters. The work includes a new method of assigning confidence to the population estimate and points out problems with methods currently available in the literature, even in the case of constant catchability. The theory is applied both to data in historical literature and to more recent data from streams in New Brunswick, Canada. These examples demonstrate that the assumption of constant catchability can frequently lead to serious errors in data interpretation. In some cases, the conclusion that the population size is well known may be blatantly false and reasonable estimates may be impossible without further data. Schnute 1983. Can. J. Fish. Aqua. Sci. 40: 2153-2169.

A manual for easy nonlinear parameter estimation in fishery research with interactive microcomputer programs

Jon Schnute

Pacific Biological Station, Nanaimo, B.C. V9R 5K6

Abstract

This manual describes in detail how to solve many practical problems encountered in nonlinear parameter estimation. In addition, it presents software to aid the user with three tasks: (1) finding optimal parameter estimates, (2) plotting observations and model predictions, and (3) displaying graphically the variation in likelihood (or sum of squares) when the parameters are varied from their optimal estimates. This software is coded in BASIC for the Apple II microcomputer, and it is available on a suitable 5 ¼” diskette. In many cases, the user can adapt the general software here to his or her particular problem by adding just a few lines of BASIC code.

The simplex method of searching for a function minimum lies at the heart of the discussion here. This manual describes the method completely. Although a simplex search is know to be leas efficient of compute time than derivative-based methods, it has the considerable advantage of minimizing human time required for coding a particular problem. The manual places great emphasis on ease of program development, even at the expense of computer time. It also shows how to adjust the simplex methods for optimal efficiency and how to apply it in other contexts besides nonlinear estimation, the discussion through is illustrated with numerous examples from fisheries literature, although the methods have obvious application in many fields.

The broad purpose of this manual is to make the reader as comfortable with the process of nonlinear estimation as with the much simpler standard procedure of linear regression. Schnute 1982. Dept. of Fish. and Oceans Fish. Res. Br. Can. Tech. Rep. Fish. Aquat. Sci. Res. Rep. 1140.

Key words

Parametric models, parameter estimation, nonlinear estimation minimization, function minimization, simplex search, microcomputers, Apple II microcomputer, BASIC, likelihood, maximum likelihood, fisheries models

A versatile growth model with statistically stable parameters

Jon Schnute

Department of Fisheries and Oceans, Resource Services Branch, Pacific Biological Station, Nunaimo, B.C. V9R 5K6

Abstract

This paper presents a new comprehensive growth model which includes numerous historical models as special cases. The new model is derived from a concise biological principle which, unlike earlier theories, relates to growth acceleration. Properties of growth curves, such as asymptotic limits or inflection points, are shown to be incidental in this new context. Possible submodels include not only asymptotic growth (such as von Bertalanffy, Richards, Gompertz, or logistic growth) but also linear, quadratic, or exponential growth. By simple analysis of variance, the observed data can be used directly in deciding which type of model is most appropriate. The new model is cast in terms of parameters which have stable statistical estimates. From this perspective, it is shown how earlier formulations sometimes result in an endless computer search for optimal parameter estimates. Schnute 1981. Can. J. Fish. Aqua. Sci. 38: 1128-1140.

A new approach to length-frequency analysis: growth structure

Jon Schnute, David Fournier

Department of Fisheries and Oceans, Resource Service Branch, Pacific Biological Station, Nanaimo, B.C. V9R 5k6

Abstract

This paper presents a new approach to length-frequency analysis which takes account of biological structure in the mean lengths and standard deviations in length for various ages-classes of fish. The new methods help determine biologically meaningful solutions, even when earlier methods lead to an ambiguous set of competing solutions. The structure of the standard deviations turns out to be especially important. For describing the means, new parameters are defined for von Bertalanffy growth which prove to have greater biological meaning and numerical stability than L_infinity. K, and t₀these new parameters can often be estimated easily from the raw data in cases where the species experiences a slowing of growth with age. This paper also presents x² methods which can be used to rank competing solutions, although the results are not definitive. All methods are illustrated using data previously published for pike and abalone. An appendix describes in detail the computer programs required for the analysis. Schnute and Fournier. 1980. Can. J. Fish. Aqua. Sci. 37: 1337-1351.

Key words

length-frequency analysis, aging of samples, von Bertalanffy growth, growth, maximum likelihood estimation, nonlinear estimation

The Simultaneous Determination of Optimal Thinning Schedule and Rotation for an Even-Aged Forest

Gerard F. Schreuder

Center for Quantitative Sciences, University of Washington, Seattle 98105

Abstract

A model to determine simultaneously the optimal thinning schedule and rotation is presented in two forms. If time is assumed to be continuous, the problem can be formulated in the calculus of variation form. In this form no closed form solution has been obtained. If the model is recast in dynamic programming form, a numeric solution can easily be obtained, without imposing any limitations on the form of the relevant revenue and cost functions. Thus the traditional Faustmann formula is generalized considerably. At the same time the close connection between the calculus of variations and dynamic programming is illustrated. Schreuder 1971. For. Sci. 17: 333-339.

Additional key words

Dynamic programming, rotation, stand density, stand growth, costs and returns.

Understanding the efficiency of model-dependant procedures with the weighted regression estimator

Hans T. Schreuder, Zhao Ouyang, H.G. Li

Rocky Mountain Forest and Range Experiment Station, USDA Forest Service, Fort Collins, Colorado, Colorado State University, For Collins, Colorado

Abstract

Eight sampling strategies using the weighted regression estimator were studied to understand why one sampling design may be more efficient than another. The results did not produce any easy explanation. A design that minimizes part of the expected mean square error is very inefficient on another part of the expected mean square error. For example, psR is, respectively, the least and most efficient on the two parts of the expected mean square error but least efficient overa1l. The extremes method is most efficient on one part, average on the other, and clearly best overall in terms of expected mean square error. Although a clear understanding is still lacking as t o why one method is more efficient than another, this exercise has been quite fruitful in suggesting promising new sampling designs.

The jackknife - a useful statistical tool

Hans T. Schreuder, G.E. Brink

Rocky Mountain Forest and Range Experiment Station, USDA Forest Service, 240 West Prospect, Fort Collins, CO 80526 (U.S.A.)

Abstract

Volume estimation by diameter class using bootstrap variance analysis

Hans T. Schreuder, G.E. Brink

Rocky Mountain Forest and Range Experiment Station, USDA Forest Service, 240 West Prospect, Fort Collins, CO 80526 (U.S.A.)

Abstract

Alternative estimators for Point-Poisson sampling

Hans T. Schreuder, G.E. Brink

Rocky Mountain Forest and Range Experiment Station, USDA Forest Service, 240 West Prospect, Fort Collins, CO 80526 (U.S.A.)

R. L. Wilson

Regional Timber Office, Container Corporation of America, P.O. Box T, Fernanaidan Beach, FL 32034

Abstract

Additional key words

Jackknife, Raj estimator, negative variance estimates, pps with and without replacement, 3p

Yield prediction for unthinned natural slash pine stands

H.T. Schreuder

Southeastern Forest Experiment Station, U.S. Forest Service, Research Triangle Park, North Carolina

W.L. Hafley

School of Forest Resources, North Carolina State University, Raleigh, North Carolina

F.A. Bennett

Abstract

Additional key words

Weibull distribution, functional models, diameter distribution, height distribution, volume distribution

A Useful Bivariate Distribution for Describing Stand Structure of Tree Heights and Diameters

Hans T. Schreuder

USDA Forest Service, Southeastern Forest Experiment Station, Forestry Sciences Laboratory, Research Triangle Park, North Carolina 21109, U.S.A

W. L. Hafley

School of Forest Resources, North Carolina State University, Raleigh, North Carolina 27607, U.S.A

Abstract

Hafley and Schreuder (1977) have shown that the marginal S_B distribution fits diameter and height data consistently better than the Weibull, beta, gamma, lognormal, and normal distributions. The bivariate extension of the S_B distribution, the S_BB is both more realistic and provides more usable information than the currently accepted approach in describing even-aged forest stand height-diameter data. The S_BB allows for the generation of bivariate frequencies for diameter and height whereas the current approach only provides marginal frequencies /or diameter. In addition, the S_BB implies a new height-diameter relationship which is comparable in fit to the most commonly used height-diameter regression model. Application of the SBB to two data sets is presented. Schreuder and Hafley. 1977. Biometrics. 33: 471-478.

Coniferous stand characterized with the Weibull distribution

Hans T. Schreuder, Wayne T. Swank

USDA Forest Service, southeastern Forest Experiment Station, Asheville, North Carolina 28802

Abstract

The Weibull distribution F (x)=1-exp[-(x/a)^c], summarized diameter, basal area, surface area, biomass, and crown profile distribution data well for several different ages of white and loblolly pine plantations. The data for diameter, basal area, surface area, and biomass were easily summarized by this one distribution in a theoretically consistent fashion. This is not possible with the normal and the gamma distribution, and the lognormal gives less satisfactory results. The distribution function should prove useful in the above variables. The change in these parameters may be a good way to characterize and interpret changes in stands over time. Schreuder and Swank. 1974. Can. J. For. Res. 4: 518-523.

An age-independent basal area increment model for maritime pine trees in northwestern Spain

Jorg Schroder

Institut fur forsteinrichtung und Ertragskunde, Georg-August-Universitat Gottingen, Busgenweg 5, 37077 Gottingen, Germany

Roque Rodriquez Soalleiro

Escuela Politecnica Superior de Lugo, Universidad de Santiago de Compostela, Campus Universitario, 27002 Lugo, Spain

Guillermo Vega Alonso

Xunta de Galicia, Delegacion Provicidal de Medio Ambiente, Calle Benito Corbal 47, 36071 Pontevedra, Spain

Abstract
An individual tree basal area increment model for pure stands of maritime pine (Pinus pinaster Ait.) is presented. The data used to fit the model were collected from 67 temporary 0.05 ha plots installed in regularly stocked, even- and uneven-aged stands located in Galicia (northwestern Spain). Explanatory variables used to predict the basal area increment are breast-height diameter outside bark, a diameter-squared term, a distance-independent competition index, crown spread ratio, and effective soil depth. Age and site index are intentionally omitted to obtain generality for a wide range of even- and uneven-aged stand conditions. The resulting model explains more than 78% of the total growth variation, corresponds to current hypotheses about tree growth, and is expected to provide reliable site-specific predictions for most of the pure maritime pine stands found in the study area. Schroder and Salleiro et al. 2002. For. Ecol. Manage. 157: 55-64.

Author Keywords
Pinus pinaster Ait., simulation, growth modeling, basal area increment

Key Words Plus
growth-model, individual trees, stands, yield, forests

Gibbs random field models: a toolbox for spatial information extraction

M. Schroder

Swiss Federal Institute of Technology ETH Zurich, Communication Technology Laboratory, Gloriastr. 35, Ch-8092, Zurich, Switzerland

M. Walessa

German Aerospace Center DLR, German Remote Sensing Data Center DFD, Oberpfaffenhofen, D-82234, Webting, Germany

H. Rehrauer

Swiss Federal Institute of Technology ETH Zurich, Communication Technology Laboratory, Gloriastr. 35, Ch-8092, Zurich, Switzerland

K. Seidel

Swiss Federal Institute of Technology ETH Zurich, Communication Technology Laboratory, Gloriastr. 35, Ch-8092, Zurich, Switzerland

M. Datcu

German Aerospace Center DLR, German Remote Sensing Data Center DFD, Oberpfaffenhofen, D-82234, WeBting, Germany

Abstract
In this paper, we present Gibbs random field models in the form of a powerful toolbox for spatial information extraction from remote sensing images. These models are defined via parametrised energy functions that characterise local interactions between neighbouring pixels. After shortly revisiting the information theoretical concept and defining a family of Gibbs models, we give a tour through examples of different kinds of spatial information extraction. These examples range from parameter estimation and analysis, via selection of the model that best describes the image data, up to the segmentation of the whole image into regions with uniform properties of the model. Finally, the concept of across-image segmentation of spatial information leads to an application for content-based queries from remote sensing image archives. Schroder and Walessa et al. 2000. Comput. Geosci. 26: 423-432.

Author Keywords
remote sensing, image processing, image segmentation

Key Words Plus
remote-sensing images, markov random-fields, texture segmentation, retrieval

A comparison of large-scale spatial vegetation patterns following clearcuts and fires in Ontario’s boreal forests

David Schroeder, Ajith H. Perera

Forest Landscape Ecology Program, Ontario Forest Research Institute, Ontario Ministry of Natural Resources, 1235 Queen St. E., Sault Sainte Marie, Ont., Canada P6A 2E5

Abstract
The role of wildfires as the most significant source of disturbance in boreal forests has been equaled by clearcuts during the past five decades. Post-disturbance revegetation patterns are important because they have a direct influence on many ecological processes. However, the knowledge of post-disturbance changes in spatial patterns of forest cover is scarce, especially at large scales.

We examined spatial patterns of forest cover in a four decade series of post-fire and post-clearcut landscapes in boreal Canada. A suite of indices was used to quantify spatial patterns of post-disturbance vegetation, based on Landsat TM imagery, and edaphic conditions. Indices were grouped in terms of patch geometry, contagion and composition. We used a general linear model to compare the effects of disturbance type, time since disturbance, edaphic conditions, and their interactions on these indices.

Clearcuts produced more heterogeneous landscapes after disturbances in comparison to fires. Time since disturbance also had a significant effect on spatial patterns of vegetation: the older disturbances had more landcover types with higher interspersion. Edaphic conditions also significantly affected spatial patterns of vegetation. Landscapes with complex spatial patterns of edaphic conditions also had complex spatial patterns of vegetation. (C) 2002 Elsevier Science B.V. All rights reserved. Schroeder and Perera. 2002. For. Ecol. Manage. 159: 217-230.

Author Keywords
boreal forests, spatial vegetation patterns, fire, clearcut

Key Words Plus
landscape, disturbance, succession, simulation, wildfire, regimes, river

Loblolly pine - the ecology and culture of loblolly pine (Pinus taeda l.)

Robert P. Schultz

U.S. Department of Agriculture Forest Service Southern Forest Experiment Station New Orleans, Louisiana

Abstract

Loblolly pine ranks as a highly valuable tree for its pulp, paper, and lumber products. In the South, loblolly is planted more than any other conifer. Loblolly Pine: The Ecology and Culture of Loblolly Pine (Pinus taeda L.) adds to the technical foundations laid by Ashe (1915) and Wahlenberg (1960). Agriculture Handbook 713 encompasses genetics, tree improvement, field inventory and analysis, and international forestry, as well as ecology, direct seeding, and planting.Loblolly Pine: The Ecology and Culture of Loblolly Pine (Pinus taeda L.) highlights individual tree, stand, and land management alternatives useful to resource managers, students, researchers, and others. Schultz 1997. USDA For. Serv. Souther FES. Agric. Handb. 713. 514p.

A computer program to evaluate timber production investments under uncertainty

Dennis L. Schwetzer

Pacific Northwest Forest & Range Experiment Station, Portland

Abstract

A computer program has been written in Fortran IV to calculate probability distributions of present worths of investments in timber production. Inputs can include both point and probabilistic estimates of future costs, prices, and yields. Distributions of rates of return can also be constructed. Schweitzer 1968. USDA For. Serv. No. Cent. For. Exp. Sta. Res. Note. NC-65.

Forest inventory and analysis needs for a growth processor

Charles T. Scott

USDA Forest Service, Northeastern Forest Experiment Station, 370 Reed Road, Broomall, PA, 19008

Jerold T. Hahn

North Central Forest Experiment, 1992 Folwell Avenue, St. Paul, MN, 55108

Colin D. MacLean

Pacific Northwest Forest and Range Experiment Station, P.O. Box, Portland, OR, 97208, USA

Abstract

The Forest Inventory and Analysis units at the experiment stations within the USDA Forest Service conduct periodic surveys of the forest lands of the United States. The FIA units must then project the forest resource; they must have appropriate growth projection modelling systems for all species and stand conditions. Currently, no such system or collection of systems exists that covers the full range of conditions encountered. The growth processor should be applicable at the state or regional level and at the individual stand level. The growth processor also should be an individual- tree , distance-independent system containing several subsystems including growth, mortality , removals, regeneration, and non-tree vegetation.

Derivation of a competitive index for individual trees from seasonal growth patterns

Frederick W. Smith

Department of Forest and Wood Sciences, Colorado State University, Fort Collins, CO, U.S.A 80523

David R. M. Scott

College of Forest Resources, University of Washington, Seattle, WA, U.S.A. 98105

Abstract

A competitive index for lodgepole pine (Pinus contorta Dougl.) trees in central Oregon is developed from seasonal basal area growth and an indirect estimator of foliar leaf area. Differences in seasonal basal area growth and the ratio of basal area growth to sapwood basal area between trees with and without neighbors are used to document growth reductions owing to the proximity of competing individuals. A regression between basal area growth and sapwood basal area (an estimator of leaf area) is used as a predictor of maximum potential basal area growth for trees growing free of competition. The competitive index is determined as the ratio of actual to potential basal area growth for individual trees. This index standardizes growth against differences in tree size and site conditions. Plant and soil water relations are considered as possible mechanisms of competitive interaction. Moderate minimum seasonal values of predawn leaf pressure potentials (- 0.76 to -0.92 MPa) and minor differences between trees in different competitive classes led to the conclusion that soil water may not be the primary mechanism of competition on this site. Smith and Scott. 1984. Can. J. For. Res. 14: 266-270.

A Data Structure Analogue for Modelling Tree Crowns

Stephen Smith, K. Scoullar

Faculty of Forestry, University of British Columbia, Vancouver, British Columbia

Abstract

A rooted-tree computer data structure is used as an analogue of coniferous tree crowns in building a simulation model of a single tree. Submodels of branch death and the allocation of stored reserves are used to illustrate the model. The programming language used is PL/I. Smith and Scoullar. 1975. Can. J. For. Res. 5: 574-579.

A consistent system of equations for tree and stand volume

Mahadev Sharma, Richard G. Oderwald, Ralph L. Amateis

Department of Forestry, Virginia Tech, Blacksburg, VA 24061, USA

Abstract
A system of equations for tree and stand volume was derived using dimensional analysis techniques. The equations are analytically compatible and numerically consistent. Two parameters define the entire system, which can be estimated by fitting the tree taper and volume equations or by fitting the stand level equations. Data from a thinning study in loblolly pine (Pinus taeda L.) plantations established on cutover site-prepared lands were used to test the utility of the system. The equations are general and can be applied to other tree species in other locales. (C) 2002 Elsevier Science B.V. All rights reserved. Sharma and Oderwald et al. 2002. For. Ecol. Manage. 165: 183-191.

Author Keywords
nonlinear least squares regression, mathematically/algebraically consistent, quadratic mean diameter, lorey mean height, dimensional analysis, loblolly pine, Pinus taeda

Key Words Plus
compatible volume, models, taper

Stem Analysis of Jack Pine (Pinus Banksiana, Lamb.): Techniques and Concepts

S. R. Shea, K. A. Armson

Faculty of Forestry, University of Toronto, Toronto, Ontario

Abstract

Detailed stem analysis data are presented for 6 trees which ranged in age from 44 to 68 years and grew on a range of soils which varied primarily in soil texture. The use of measures of specific growth such as current annual height increment, Type 3 ring width sequences and specific volume increment is discussed in relation to supply of raw materials to the tree. The change in form of tree crown, as reflected by Type 1 ring width sequences is related to the other measures of specific growth. The use of selected parameters of tree growth in relation to tree development and its possible application in silviculture is discussed. Shea and Armson. 1972. Can. J. For. Res. 2: 392-406.

Forest statistics for southwest Georgia, 1996

Raymond M. Sheffield, Michael T. Thompson

Forest Inventory and Analysis group, Southern Research Station, U.S. Department of Agriculture, Forest Service, Asheville, NC.

Abstract

This report highlights the principal findings of the seventh forest survey of Southwest Georgia. Field work began in June 1995 and was completed in November 1995. Six previous surveys, completed in 1934, 1951, 1960, 1971, 1981, and 1988 provide statistics for measuring changes and trends over the past 62 years. This report primarily emphasizes the changes and trends since 1988. Periodic surveys of forest resources are authorized by the Forest and Rangeland Renewable Resources Research Act of 1978. These surveys are a continuing, nationwide undertaking by the Regional Experiment Stations of the U.S. Department of Agriculture, Forest Service. In the Southern United States, these surveys are conducted by the Forest Inventory and Analysis (FIA) Research Work Unit at the Southern Research Station, Asheville, NC. The FIA unit operates out of two locations, one in Starkville, MS, and the other in Asheville, NC, and is responsible for inventories of 13 Southern States and the Commonwealth of Puerto Rico. The primary objective of these surveys is to periodically inventory and evaluate all forest and related resources. These multiresource data help provide a basis for formulating forest policies and programs and for the orderly development and use of the resources. This report discusses the extent and condition of forest land, associated timber volumes, and rates of timber growth, mortality, and removals. Sheffield and Thompson. 1997. USDA For. Serv. Res. Pap. SRS-12. 68p.

Growth projection systems in forest management

J. Harry G. Smith

Forest Resources Management, Faculty of Forestry, University of British Columbia Vancouver, B.C., V6T 1W5

Abstract

Results of review of the variable density growth projection system of the Inventory Branch of the Ministry of Forests are summarized. Information and methods needed to improve protection and silviculture and to make better estimates of growth and yield for management are reviewed. Needs for new operationally sound methods are outlined for three neglected areas of management. It is concluded that existing inventory branch methods represent existing stands adequately and now can provide useful simulation estimates of future forest conditions. However much more effort should be devoted to estimating future development of very young stands, to classifying relative growth of young, mature and overmature stands and to developing better regression methods for estimating trends in stocking and density. Smith and Vissage et al. 1984. Forest. For. Chron. 60: 150-156.

Forest resources of the United States, 1997

W. Brad Smith

Science, Policy, Information, and Inventory Staff, Washington Office, U.S. Department of Agriculture, Forest Service, Washington, DC.

John L. Vissage

Forest Inventory and Analysis Unit, North Central Research Station, U.S. Department of Agriculture, Forest Service, St. Paul, MN.

David R. Darr

Resource Valuation and Use Staff. Washington Office, U.S. Department of Agriculture, Forest Service, Washington, DC

Raymond Sheffield

Resource Analysis Unit, Southern Research Station, U.S. Department of Agriculture, Forest Service, Asheville, NC

Abstract

Forest resource statistics from the 1987 Resources Planning Act (RPA) Assessment were updated to 1997 to provide current information on the Nation`s forests. Resource tables present estimates of forest area, volume, mortality, growth, removals, and timber products output in various ways, such as by ownership, region, or State. Current resource data are analyzed and trends since 1987 are noted. Resource trends are placed within the context of changes in the timber resources since 1953. A fold-out forest type map produced from satellite imagery provides a visual display of the location of forest land. Smith and Vissage et al. 2001. USDA For. Serv. Gen. Tech. Rep. NC-219.

Key words

RPA, assessment, inventory, forest statistics, area, volume, forest history, AVHRR, map.

Growth after thinning a 35-year-old natural stand to different loblolly pine and hardwood basal areas

Michael G. Shelton, Paul A. Murphy

USDA Forest Service, Southern Research Station, Monticello, AR 71656-3516

Abstract

Growth was monitored for 4 years in a thinned stand in southern Arkansas with three pine basal areas (70, 85, and 100 ft2/ac) and three hardwood basal areas (0, 15, and 30 ft2/ac); pretreatment basal areas averaged 119 and 33 ft2/ac for pines and hardwoods, respectively. Treatments were arranged in a 3 x 3 factorial randomized complete block design with three replicates, yielding 27 permanent 0.20 acre plots. Growth variables were regressed with residual pine and hardwood basal areas. Pine basal area and volume growth increased with the pine stocking level after thinning and decreased with the level of retained hardwoods. For basal area and merchantable volume, hardwood growth largely compensated for losses in the pine component, and thus, hardwood retention had little net effect on the total growth of the stand. The greatest impact of hardwood retention was on the stand's sawtimber growth, because hardwoods did not contribute to this product class. Each 1 ft2/ac of retained hardwood basal area reduced pine sawtimber growth by 6 to 10 bd ft (Doyle/ac/yr, depending on the pine stocking. Because large differences existed in the value of timber products, retaining 15 and 30 ft2/ac of hardwoods reduced the value of' timber production by 13 and 24 percent, respectively, at 4 years after thinning. Shelton and Murphy. 1997. South. J. Appl. For. 21: 168-174.

Fertilization fails to increase diameter growth of sawlog-size northern hardwoods in upper Michigan

Douglas M. Stone

North Forest Experiment Station Bedford Indiana

Stephen G. Shetron

Ford Forestry Center Michigan Technological University

James Peryam

Lake States woodlands Division American Compary, Amasa, Michigan

Abstract

Five hundred individual, sawlog-size crop trees in a selectively managed northern hardwood stand were fertilized with 2.61 kg (5.75 lbs) of N-P-K fertilizer per tree in June 1970. The stand is predominantly sugar maple (Acer saccharum Marsh.) growing on a deep silt loam soil of good site quality. Three- and six-year diameter at breast height (dbh) and basal area growth were analyzed by treatment and diameter class. Fertilization did not increase dbh or basal area growth during either measurement period. Mean annual dbh growth of both fertilized and control trees was 0.41 cm (0.16 in.); trees in all diameter classes grew at essentially the same rate. Nearly half of the study trees grew 2.54 cm (1.0 in.) or more during the 6-year period, and fertilization did not change this proportion. Natural supplies of nutrients in good hardwood soils appear adequate for rapid growth of these species. Periodic cuttings maintain superior growth rates and contributes to efficient nutrient cycling. Fertilization is not recommended on sites like this because growth responses are not likely to justify costs. Stone and Shetron et al. 1982. For. Chron. 58: 207-210.

Quick estimation of the three-parameter Weibull to describe tree size distributions

Stephen Shifley, Ellen Lentz

North Centrl Forest Experiment Station, 1992 Folwell Avenue, St. Paul, MN 55108

Abstract

The method of moments is an alternative to the maximum likelihood method of estimating values of Weibull density function parameters that describe the size distribution of trees. In contrast to maximum likelihood estimation, the method of moments can be applied when only the mean and variance of tree sizes in a sample are known. With the aid of the special tables presented, method of moments computations to estimate Weibull scale and shape parameters can be quickly carried out by hand. The ratio of the sample mean over the standard deviation is shown to be a simple index to the shape of the associated Weibull density function. Equations used in solving for Weibull parameters by the method of moments provide a direct link between parameter prediction models and parameter recovery models based on the Weibull density function. Shifley and Lentz. 1985. For. Ecol. Manage. 13: 195-203.

Determination of spatial continuity of soil lead levels in an urban residential neighborhood

Nita J. Shinn, John Bing-Canar, Michael Cailas

School of Public Health, University of Illinois at Chicago, Chicago, Illinois

Nicholas Peneff

Public Health & Safety, Inc., Chicago, Illinois

Helen J. Binns

Children’s Memorial Hospital, Northwestern University, Chicago, Illinois

Abstract
This study uses geostatistical techniques to model and estimate soil lead levels in an urban, residential neighborhood. Sixty-two composite soil samples (median 1773 ppm; range 175 to 7953 ppm) in a four-block area of brick and stone homes were obtained. The spatial continuity of soil lead levels was modeled with a semi-variogram, which was then used to estimate lead levels at unsampled locations, a process called kriging. Because soil lead levels were spatially correlated, it is likely that a "nonrandom" process generated the lead distribution found. This finding signifies the existence of lead sources which were tentatively identified on historical maps of the area and from past traffic volume patterns. The distribution of kriged estimates of soil lead levels provides an explanatory tool for exploring and identifying potential sources and mag be useful for targeting urban soil abatement efforts. Shinn and Bing-canar et al. 2000. Environmental Research Section A. 82: 46-52.

Author Keywords
soil, lead, mapping, geostatistics, urban

Key Words Plus
blood lead, children, smelter, exposure, plants

Sample Sizes and Estimation Methods for the Weibull Distribution for Unthinned Slash Pine Plantation Diameter Distributions

Barry D. Shiver

School of Forest Resources, University of Georgia, Athens, Georgia 30602.

Abstract

Three methods were evaluated over four sample sizes for prediction of diameter distributions in unthinned site prepared slash pine plantations. Maximum likelihood estimation provided the best estimates of known distribution parameters. Under the likely assumption that there are no specific underlying parameters, both modified moments estimation and a percentile estimator reproduce the underlying distribution of diameters as well or better than maximum likelihood estimation. All three methods require approximately 50 tree samples to reproduce distributions into classes with less than 10% error in any class. Shiver 1988. For. Sci. 34: 809-814.

Additional Key Words

Parameter estimation, probability density function, Pinus elliottii

Determination of stand properties in boreal and temperate forests using high-resolution imagery

Herman H. Shugart

Department of Environmental Sciences, University of Virginia, Charlottesville, VA 22901

Laura L. Bourgeau-Chavez

ERIM International Inc., P.O. Box 134008, Ann Arbor, MI 48113-4008

Eric S. Kasischke

ERIM International Inc., P.O. Box 134008, Ann Arbor, MI 48113-4008.

Abstract
The existence of a relatively long (ca. 40 yr) satellite imagery archive for examination of potential worldwide forest change motivated an inspection of the relation between forest features observable from higher resolution airborne and satellite imagery and measures of forest biomass, height, and age. Using these data, we inspected the relation between stand age, mean diameter, height, and standing aboveground biomass for forest stands located in boreal forest (near Fairbanks, Alaska) and temperate forest (near Durham, North Carolina). The features of a spatial statistic, semivariance, prove to be related to the structure, age, and biomass patterns of these forests. These initial results indicate the feasibility of using appropriately calibrated high-resolution satellite imagery to estimate and monitor aboveground carbon reserves of forests. Shugart and LauraBourgeau_et al. 2000. For. Sci. 46: 478-486.

Author Keywords
semivariance, Alaska, North Carolina, remote sensing, biomass

Key Words Plus
climate change, biomass

Site quality estimation using height and diameter

Benjamin B. Stout

School of Forestry, University of Montana, Missoula MT 59812

Durland L. Shumway

Concrete, WA 98237

Abstract

A method for estimating site quality using height and diameter of dominant and codominant trees independent of tree age is examined. Eastern hardwoods are the test species. Two parameters define the H-D site curves: a site parameter and a species parameter. The estimated site parameters increase with increasing site index and are significantly different between site classes. The estimated species parameters are species specific and are not significantly different between site classes. Curves for estimating site class are presented. Some species reach a higher proportion of total height at a given diameter. The effects of these apparent height growth patterns are discussed from a silvicultural viewpoint. Stout and Shumway. 1982. For. Sci. 28: 639-645.

Additional Key words

Site index, hardwoods

Process-based models and forest management

Risto Sievanen

Department of Mathematics, The Finnish Forest Research Institute Unioninkatu 40 A, 00170 Helsinki Finland

Thomas E. Burk

Department of Forest Resources, University of Minnesota 1530 N Cleveland Avenue, St. Paul, MN 55108 USA

Abstract

Construction of a stand growth model utilizing photosynthesis and respiration relationships in individual trees

Risto Sievanen

Department of Mathematics, The Finnish Forest Research Institute Unioninkatu 40 A, 00170 Helsinki Finland

Thomas E. Burk, Alan R. EK

Department of Forest Resources, University of Minnesta 1530 N Cleveland Avenue, St. Paul, MN 55108 USA

Abstract

A discrete growth model for an individual tree in a forest stand is presented. The model is based on simplified photosynthesis and respiration relationships. Implementation of the model also requires functions for change in live crown ratio and stems per unit area. Tree growth predictions are expressed in terms of biomass and basal area. A preliminary evaluation of the model is made using data from the literature on the growth of even-aged stands of red pine (Pinus resinosa Ait.). The model produced both short- and long-term trends consistent with stand growth data. The evaluation suggests further possibilities for model component refinement. Sievanen, Burk et al.1988. Can. J. For. Res. 18: 1027-1035. 1988.

Parameter estimation in a photosynthesis-based growth model

Risto Sievänen

The Finnish Forest Research Institute, Department of Mathematics, Unioninkatu 40 A, 00170 Hlesinki, Finland

Thomas E. Burk, Alan R. Ek

Department of Forest Resources, University of Minnesota, 1530 North Cleveland Avenue, St. Paul, MN 55108

Abstract

A growth model for an individual tree in a forest stand based on photosynthesis and respiration relationships is briefly described. Individual tree growth data (six measurements of diameter) of red pine (Pinus resinosa Ait.) stands are used to estimate a subset of the parameters in the model using a globally, simultaneous approach. The loss function is composed of prediction errors for individual trees at the times of measurement and is a nonlinear function of the parameters. Estimates for parameters are compared with values found in the literature. Estimation suggests ways the structure of the model should be changed. Use of alternative loss function definitions provides insight into model component stability.

A comparison of two parameter prediction methods for stand structure in Finland

Jouni Siipilehto

Finnish Forest Research Institute/Vantaa Research Centre, P.O. Box 18, FIN-01301 Vantaa, Finland

Abstract
The objective of this paper was to predict a model for describing stand structure of tree heights (h) and diameters at breast height (dbh). The research material consisted of data collected from 64 stands of Norway spruce (Picea abies Karst.) and 91 stands of Scots Dine (Pinus sylvestris L.) located in southern Finland. Both stand types contained birch (Betula pendula Roth and B. pubescent Ehrh.) admixtures. The traditional univariate approach (Model I) of using the dbh distribution (Johnson's S-B) together with a height curve (Naslund's function) was compared against, the bivariate approaches, Johnson's S-BB distribution (Model II) and Model I-epsilon. In Model I-epsilon within-dbh-class h-variation was included by transforming a normally distributed homogenous error of linearized Naslund's function to concern real heights. Basal-area-weighted distributions were estimated using the maximum likelihood (ML) method. Species-specific prediction models were derived using linear regression analysis. The models were compared with Kolmogorov-Smirnov tests for marginal distributions, accuracy of stand variables and the dbh-h relationship of individual trees. The differences in the stand characteristics between the models were marginal. Model I gave a slightly better fit for spruce, but Model II was better for pine stands. The univariate Model I resulted in clearly too narrow marginal h-distribution for pine. It is recommended applying of a constrained ML method for reasonable dbh-h relationship instead of using a pure ML method when fitting the S-BB model.

Author Keywords
parameter prediction, dbh and height distribution, Johnson's S-BB distribution, Naslund's height curve, Picea abies, Pinus sylvestris

Key Words Plus
diameter distributions, bivariate distribution, heights

Current biogeoclimatic ecosystem classification in the Bow-Crow forest: a framework for management and productivity assessment

Bela Sivak

Forest Research Branch Alberta Forest Service

Abstract

The ongoing forest site classification program of the Forest Research Branch, AFS, identifies forest sites (Plant Associations) for operational use based on plant-available moisture and nutrients, within zones homogeneous in terms of climate, vegetation and soil features. The levels o f moisture and nutrients for each site are determined on the basis of measured site factors including features of climate, soils, landforms, and vegetation. This ecosystem approach and its emphasis on plant-available moisture and nutrients provides a suitable framework to manage and use forest ecosystems according to their capability. Sivak 1988.

The impact of grazing on spatial variability of humus profile properties in a grass-encroached Scots pine ecosystem

Annemieke Smit

Department of Physical Geography and Soil Science, University of Amsterdam, The Netherlands Centre for Geo-Ecological Research, Nieuwe Prinsengracht 130, 1018 VZ Amsterdam, Netherlands

Abstract
Grazing is often used in management of grass-encroached Scots pine (Pinus sylvestris) forests in The Netherlands, because it is thought to increase diversity of vegetation and humus profiles. To test this hypothesis, two 1 ha plots in a primary pine stand on drift sand were grazed for 7 years at high and moderate intensities. The humus profiles of these two sites and of a control site were sampled at various scales to assess changes in grass biomass, thickness of ectorganic horizons, organic matter stocks, root stocks, and pH. Whether grazed or not, site factors such as slope angle and aspect, distance to trees and parent material did not explain the spatial variability of the plots. Spatial variability could be linked to the tussock structure of the grass layer. With continued grazing, ectorganic horizons became more compact (particularly the F horizon), stocks of organic matter increased, grass biomass decreased, variances tended to decrease and ranges tended to increase. These results suggest that at the landscape scale, where much variation in grazing intensity can be expected, spatial variability of humus profile properties and thus site diversity will increase. However, grazing intensities in the experiment were far greater than those normally employed in management of grass-encroached forests. Therefore grazing will scarcely affect site diversity, unless grazing variability is enhanced by special treatments, such as temporary fencing to create locally high grazing intensities. Smit, 1999. Catena. 36: 85-98.

Author Keywords
spatial variability, humus form, grazing, organic matter, sampling strategy

Key Words Plus
forest floor, soil

Derivation of a competitive index for individual trees from seasonal growth patterns

Frederich W. Smith

Department of Forest and Wood Sciences, Colorado State University, Fort Collins, CO, USA 80523

David R. M. Scott

College of Forest Resources, University Washington, Seattle, WA, USA 98105

Abstract

A competitive index for lodgepole pine (Pinus Contorta Dougl.) trees in central Oregon is developed from seasonal basal area growth and an indirect estimator of foliar leaf area. Differences in seasonal basal area growth and the ratio of basal area growth to sapwood basal area between trees with and without neighbors are used to document growth reductions owing to the proximity of competing individuals. A regression between basal area growth and sapwood basal area (an estimator of leaf area) is used as a predictor of maximum potential basal area growth for individual trees. This index standardizes growth against differences in tree size and site conditions. Plant and soil water relations are considered as possible mechanisms of competitive interaction. Moderate minimum seasonal values of predawn leaf pressure potentials (-0.76 to –0.92 MPa) and minor differences between trees in different competitive classes led to the conclusion that soil water may not be the primary mechanism of competition on this site. Smith and Scott. 1984. Can. J. For. Res. 14: 266-270.

Climate Change and Carbon Pools in a Semi-Arid Ecosystem

Jeffrey L. Smith

USDA-Agricultural Research Service, 215 Johnson Hall, Washington State University, Pullman, WA, 99164-6421, USA

Harvey Bolton Jr.

Environmental Microbiology Group, Pacific Northwest National Laboratory, Mailstop K4-06, 900 Battelle Blvd., Richland WA 99352, USA

Abstract

We investigated the effects of climate on C fractionation and storage over a 500m elevation transect. The transect parent material, loess, was deposited at the same time, the climate has been stable for the last 3,000 y, and the vegetation is similar. Thus, the variables over the transect are mean maximum temperature ranging from 30°C at the lower slope to 24°C at the upper slope, and precipitation ranging from 180 mm at the lower slope to 270 mm at the upper slope. Soil, sampled under grass and in bare areas, was analyzed for C in the POM (particulate organic matter) fraction and in the fine mineral fraction. In both soil types, there was a significant increase in POM-C with elevation, more so in soil associated with grass. The trends for POM-N follow the POM-C with highly significant differences across the transect and between the soil type. The difference in climate from 300 m and 800-m elevation has influenced the C and N content of the soil, both in bare areas and under grass. The increase in plant biomass and decrease in decomposition is the cause of the increase in soil C and N over the transect.

Key words

climate change, soil carbon, particulate organic matter

Growth projection systems in forest management

J. Harry G. Smith

Forest Resources Management, Faculty of Forestry, University of British Columbia Vancouver, B. C., V6T 1W5

Abstract

New non-linear models can improve estimates of growth and yield

J. Harry G. Smith, A. Kozak

Forest Resources Management, Faculty of Forestry, University of British Columbia Vancouver, B. C., V6T 1W5

Abstract

Some important recent improvements to methods for estimating growth and yield are summarized and evaluated. These new methods and concepts should be considered carefully and used cautiously by resource managers and researchers seeking better methods for estimating growth and yield. Smith and Kozak. 1984. For. Rev. 63: 41-45.

Effects of species mixtures and inventory zones on net yields of British Columbia forests

J. Harry G. Smith

Forest Resources Management, Faculty of Forestry, University of British Columbia Vancouver, B. C., V6T 1W5

Richard C. Yang

Northern Forest Research Centre, Canadian Forestry Service, Edmonton, T6H 3S2

Abstract

A least squares procedure developed to determine statistical significance of both qualitative and quantitative data was applied to find proportions of plots and amounts of wood yield by species mixtures and inventory zones. For 9781 plots in Douglas-fir, spruce, and lodgepole pine forest types, sample volumes were highest in mixtures of these species with other conifers. However, small differences in age, site index, and stand density could account for much of the variation. When stand volumes were brought to common levels of age, site and density the need to recognize species composition classes and forest inventory zones diminished.

If indicated gains from mixtures were real, stand managers could easily improve density of pure stands to exceed the increased yields indicated for conifer mixtures with interior Douglas-fir (20.4) and with lodgepole pine (3.4%). Because our analyses of regional inventories could not provide conclusive results, well designed long-term field trials would be needed to evaluate effects of species mixtures under fully comparable conditions.

When stand density is used with age and site index to estimate yields there should be no need to recognized inventory zones within Coast and interior races of the species tested. Smith and Yang. 1982. For. Chron. 58: 85-90.

Potentials for controlling bark percentage of the commercial tree species of British Columbia

J. Harry G. Smith, A. Kozak

Forest Resources Management, Faculty of Forestry, University of British Columbia Vancouver, B. C., V6T 1W5

Abstract

Diameters inside and outside bark from 13 sections were used to define variation in bark percentages for 33 844 trees representing 28 major groups of the commercial tree species of British Columbia. The range of age, height, dbh, and dbh/height associated with each group was determined. Statistical significance of effects of these factors and of up to 12 inventory zones was determined for double bark thickness as a percentage of dbh. Bole bark volumes were compared with wood volumes for trees exceeding two standard deviations of bark percentage at breast height. Thick barked trees were more common and a few have equal portions of wood and bark. Our description of the characteristics of the trees with least or most bark should help guide tree improvement programs. Identification of the extent to which stand factors can influence average bark characteristics may help timber managers grow trees of desired bark proportions. Smith and Kozak. 1981. For. Chron. August: 156-161.

Examination of methods for evaluating remining a mine waste site Part I. Geostatistical characterizaiton methodology

Martin L. Smith, Roy E. Williams

College of Mines and Earth Resources, University of Idaho, Moscow, ID 83843, USA

Abstract
Comprehensive characterization of the distribution and metal content of mine wastes is a crucial component of any project which aims at selective remediation as a function of minimization of cost and risk. Knowledge of the existence of structures and horizons within the sediments is critical to remediation since metal concentrations can vary significantly both within and between groupings of wastes and sediments. The exploratory techniques for structural and statistical analyses of the wastes as presented herein were used to identify population domains between observed waste/sediment classes and by locality within waste/sediment classes. Analysis of Variance (ANOVA) of split spoon samples extracted from the wastes proved to be effective in identifying population domains within the various observed waste/sediment classes. Variography identified the nature and continuity of variability in metal concentrations. Metal concentrations in mine and mill waste may vary rapidly over distance. For this reason, exploratory variogram analysis was used to relate estimation variance to sampling grid design and orientation. Where the deposition of mine wastes is influenced by fluvial processes, gravity separation can take place across areas that are sometimes flooded. This process produces drift in metal concentrations and call complicate the characterization of water deposited or reworked waste/sediment sites. Contaminant levels in mine and mill wastes are controlled by forces that can be very local in nature. As a result, the covariance structure can be non-stationary over relatively short distances. A methodology is presented herein for checking covariance stationarity. Smith and Williams. 1996. Eng. Geol. 43: 11-21.

Structural stage in Pacific Northwest forests estimated using simple mixing models of multispectral images

Donald E. Sabol Jr., Alan R. Gillespie, John B. Adams, Milton O. Smith

Laboratory for Terrestrial Physics, NASA Goddard Space Flight Center, Greenbelt, MD 20771, USA

Compton J. Tucker

Department of Earth and Space Sciences, Campus Box 351310, University of Washington, Seattle, WA 98195, USA

Key Words Plus
thematic mapper, tm data, canopy, reflectance, scattering, vegetation, variograms, age

A growth model based on the self-thinning rule

N.J. Smith

Department of Forest Science, University of British Columbia, Vancouver, B.C., Canada V6T 1W5

D. W. Hann

Department of Forest Management, Oregon State University, Corvallis, OR, USA, 97331

Abstract

A two staged stand growth models is developed to describe the relationship between or volume and numbers of stems in even-aged, monospecific plant populations undergoing self-thinning. The model is tested on red alder (Alnus rubra Bong) seedlings and red pine (Pinus resinosa Ait.) stands grown over a range of site qualities and initial spacings. First, survival rate is modeled as a Weibull distribution. This is then fit to an analytical size-density model to give growth estimates. Crown closure is estimated to occur at a relative density of 0.09 for red alder, while initial mortality is estimated to occur at a relative density of 0.12 for red pine. Net stand growth rates peaked at a relative density of 0.54 for red alder biomass and from relative densities from 0.40 (widest initial spacing) to 0.55 (densest initial spacing) for red pine total stem volume. Site quality merely shifted the magnitude of these relationships. The model adds a dynamic component to the self-thinning rule and also generalizers and extends the rule to stand development between crown closure and the self-thinning asymptote. Smith and Hann. 1986. Can. J. For. Res. 16: 330-334.

A new analytical model based on the -3/2 power rule of self-thinning

N.J. Smith

Department of Forest Science, University of British Columbia, Vancouver, B.C., Canada V6T 1W5

D.W. Hann

Department of Forest Management, Oregon State University, Corvallis, OR, USA 97331

Abstract

A nonlinear analytical model is developed to describe the relationship between average plant size (weight or volume) and stand density in single-aged, monospecific plant populations. The model gives estimates of the slope and intercept of the –3/2 power rule asymptote, the nature of the size-density trajectory, and such features as relative density at crown closure and the effects of soil type or site index. The model is tested by growing red alder (Alnus rubra Bong.) seedlings under greenhouse and lath house conditions at three initial spacings (8X8, 4X4, and 2X2 cm) and two soil types (river loam and alder forest soil ) for 525 growth days. There are seven harvests, starting at crown closure. All size-density trajectories tend consistently towards the same single asymptote irrespective of initial spacing , soil type, or age. The asymptote slope and intercept are 1.46 and ca. 94 kg. tree^-1. m^-2. the crown closure line is parallel to the asymptote at a relative density of 4.6X10^-3. the model also adequately describes the size-density trajectories for 20-50 year-old red pines (Pinus resinosa Ait.) growing at six initial spacings. The asymptoteslope and intercept are 1.6 and ca. 87X10³.tree^-1.ha^-1, respectively. Smith and Hann. 1984. Can. J. For. Res. 14: 656-660.

Using competitive stress index to estimate diameter growth for thinned Douglas-fir stands

S. H. Smith

Potlatch Corporation, Lewsion, Idaho

J. F. Bell

Department of Forest Management, Oregon State University, Corvallis, OR 97331

Abstract

Additional Key Words

lntertree competition, growing space, influence zone, stand models.

Asymptotic site-index curves, fact or artifact

Victor G. Smith

Faculty of Forestry, University of Toronto, Toronto, Ontario, M5S 1A1

Abstract

In the absence of permanent sample plot data, site-index curves and equations showing the height/age relationship for a tree species growing on a given site have been derived using stem analysis and temporary sample plot data. Initially the guide-curve method using temporary sample plot data (Bruce and Schumacher 1950) was employed to produced anamorphic site-index curves. More recently mensurationists (e.g. Heger 1968, and Payandeh 1978) have used stem analysis data and various regression techniques to produce polymorphic curves. This study suggests that the asymptote site-index curves (I.e., curves that tend to level off within the life of the stand) that have been produced for black spruce may be the result of using mensurational and sampling techniques in forests where site-index is correlated with age. Also the regression techniques commonly used to estimate the parameters in site-index equations do not allow for error in both the dependent variable (height) and an independent variable (site). As a consequence the site-index curves produced by these techniques are asymptotic whereas height estimates obtained from 40 black spruce permanent sample plots do not exhibit any strong asymptotic height properties for stand up 180 years of age. Smith 1984. For. Chron. 60: 150-156.

Compatible basal area growth and yield models consistent with forest growth theory

Victor G. Smith

Faculty of Forestry, University of Toronto, Toronto, Ontario M5S 1A1, Canada

Abstract

Compatible growth and yield models often include constraints which prevent them from the sample data as well as empirically derived equations. However, the superiority of such constrained equations can be demonstrated by examining their fit to the complementary stand characteristic. For example, if parameter estimation has been performed using growth as the dependent variable, the compatible yield equations should be evaluated as part of the process of choosing the best model. Thus chances are reduced of either selecting a model that has been strongly influenced by unusual aberrations in the sample data or selecting a model based on inappropriate constraints.

To illustrate, one empirical and two constrained basal area growth models, have been fitted to a common data set obtained form red pine (Pinus resinosa Ait.) plantations that were subject to a variety of thinning regimes.

One constrained model in which an inverse relationship between growth and current basal area was specified, gave the poorest fit to the growth data used for parameter estimation. However, the compatible yield form estimated future yields the best of all three models. This anomalous result provides some justification for using the new model for growth and yield estimation. Smith 1983. For. Sci. 29: 279-288.

Additional Key Words

Pinus resinosa, growth models, stand growth, thinning, basal area increment, yield estimates, density, constrained models.

The use of product form in form-class volume equations

V.G. Smith

Faculty of Forestry and Landscape Architecture, University of Toronto, Toronto, Ontario M5S 1A1

Abstract

This paper describes studies on testing of form-class volume equations in tree-length logging operations where upper-tree diameters are easily measured, and it investigates a modified procedure for estimating the relationship between diameter and the product of form and length. Stacked wood measurements for some 84 piles of spruce-fir tree lengths, obtained on four different logging operations in Newfoundland and Labrador in 1973-1974, were compared with corresponding estimates obtained by using regional volume, standard volume, form-class volume, and product-form volume equations. Product-form volume equations gave more accurate estimates than any of the other equations for these data. Regional volume/diameter equations were more accurate than the standard volume equations and were on a par with the conventional form-class volume equations. Regional volume/diameter equations should be used when the cost of developing local volume tables from sample observations is too great. Smith 1975. Can. J. For. Res. 6: 93-103.

Forest tree production centres in Canada 1983

J.H. Smyth, A.J. Brownwright

Great Lakes Forest Research Center, Canadian Forestry Service, Department of the Environment

Abstract

A total of 103 public sector, private and forest industry nurseries produced planting stock for reforestation purposes in 1983. Of these, 47 grew bare-root stock and 83 grew containerized seedlings. Planned production of bare-root and containerized stock for 1983,was 281 million and 260 million seedlings, respectively, Since 1979, bare-root production expanded by 40%, or 81 million seedlings, and containerized stock increased by 140%, or 152 million seedlings. Total production is forecasted to increase to 285 million bare-root and 271 million containerized seedlings in 1984. Smyth and Brownwright. 1984. Can. For. Serv. Dept. Env.,. Inf. Rep. O-X-357.

A ratio estimator for bias correction in logarithmic regressions

P. Snowdon

Division of Forestry and Forest Products, Commonwealth and Scientific Research Organization, P.O. Box 4008, Queen Victoria Terrace, ACT 2600, Australia

Abstract
It is recommended that the proportional bias in logarithmic regressions be estimated from the ratio of the arithmetic sample mean and the mean of the back-transformed predicted values from the regression. Under the assumption of a lognormal distribution of errors, the conditions for application of this ratio estimator are optimal. A simulated sampling study has shown that this method gives more reliable results than the methods recommended by Baskerville (G.L. Baskerville. 1972. Can. J. For. Res. 2: 49-53) or that derived by Finney (D.J. Finney. 1941. J. R. Stat. Soc. 7(Suppl.): 55-61). The new method is also less sensitive to departures from the assumption of a lognormal distribution than the other two methods. Snowdon 1991. Can. J. For. Res. 21: 720-724.

Key Words Plus
pinus-radiata, biomass, transformation

An age-independent basal area increment model for maritime pine trees in northwestern Spain

Schroder J, Soalleiro RR, Alonso GV

Abstract
An individual tree basal area increment model for pure stands of maritime pine (Pinus pinaster Ait.) is presented. The data used to fit the model were collected from 67 temporary 0.05 ha plots installed in regularly stocked, even- and uneven-aged stands located in Galicia (northwestern Spain). Explanatory variables used to predict the basal area increment are breast-height diameter outside bark, a diameter-squared term, a distance-independent competition index, crown spread ratio, and effective soil depth. Age and site index are intentionally omitted to obtain generality for a wide range of even- and uneven-aged stand conditions. The resulting model explains more than 78% of the total growth variation, corresponds to current hypotheses about tree growth, and is expected to provide reliable site-specific predictions for most of the pure maritime pine stands found in the study area. Schroder and Soalleiro et al. 2002. For. Ecol. Manage. 157: 55-64.

Author Keywords
Pinus pinaster Ait., simulation, growth modeling, basal area increment

Key Words Plus
growth-model, individual trees, stands, yield, forests

Height-diameter equation for first rotation eucalypt plantations in Portugal

Paula Soares, Margarida Tome

Department of Forestry, Instituto Superior de Agronomia, Universidade Tecnica de Lisboa, Tapada da Ajuda, 1349-017 Lisbon, Portugal

Abstract
A height-diameter equation for eucalypt plantations was developed based on a tree dataset from trials and permanent plots located in the north and central coastal regions of Portugal. The total dataset was split into two datasets through restricted random sampling at the plot level. The equations selected in one data subset were evaluated with the other subset and vice versa. Harrison equation, fitted with the iteratively reweighted least squares method, in both versions-restricted and not restricted to pass through the point diameter-height (0, 1.30)-was selected. The first version was recommended for young plantations; it is age dependent and requires a measure of stand productivity. The second version was appropriate to use in commercial forest inventory where trees smaller than 4 cm diameter are not measured; it is age independent, density dependent and, also requires a measure of stand productivity. Soares and Tome. 2002. For. Ecol. Manage. 166: 99-109.

Author Keywords
tree height prediction, height-diameter equation, plantations, Eucalyptus globulus Labill

Key Words Plus
inventory data, models, prediction, pine

A growth model of natural and silviculturally treated stands of even-aged northern hardwoods

Dale S. Solomon

Northeastern Forest Experiment Station ,USDA Forest Service

Abstract

A computer programming system was developed which simulates the development and treatment of even-aged northern hardwoods in New England. The sampling phase begins with species distributions by quality classes. Natural mortality rates are applied, and a weeding option is available to select crop trees. A diameter distribution is generated for each species, and the smaller tree in the understory are dropped from the stand. The poletimber-rotation age or diameter, using available staocking guides, yield-table data, and gross growth estimates. Thinning and final havest yields are presented by species and quality classes, using available data, the connecting phases of the simulator have been tested to determine the effects of silvicultural treatments ( or lack of treatments) on long-term stand response. Solomon 1977. Ne. For. Exp. Stn., USDA For. Serv., Broomall, Pa. Tech. Rep. NE-36.

Allometry and biomass of Korean pine(Pinus koraiensis) in central Korea

Yowha Son

Department of Forest Resources and Environmental Sciences, Korea University, Seoul 136-701, South Korea

Jae Woo Hwang

Department of Forest Resources, Yeungnam Univesity, Kyungsan 712-749, South Korea

Zin Suh kim, Woo Kyun Lee, Jong Sung Kim

Department of Forest Resources and Environmental Sciences, Korea University, Seoul 136-701, South Korea

Abstract
Aboveground tree biomass of Korean pine (Pinus koraiensis Sieb. et Zucc.) was determined for a natural forest of Korean pine and mixed deciduous trees and seven age classes of plantation forests in central Korea. Regression analyses of the dry weights of stem wood, stem bark, branches, and needles versus diameter at breast height were used to calculate regression equations of the form of log Y = a + b log X. Biomass of Korean pine in the mixed forest was 118 Mg ha(-1), and biomass in the plantations was linearly related to stand age, ranging from 52.3 Mg ha(-1) in 11 to 20-year-old stands to 317.9 Mg ha(-1) in 71 to 80-year-old stands. The proportions of stem wood and stem bark in the total aboveground biomass decreased with stand age while those of branch and needle increased. Specific leaf area of Korean pine ranging from 35.2 to 52.1 cm(2) g(-1) was significantly different among crown positions and needle ages; in general, lower crown position and current needles had the greatest surface area per unit dry weight. Son and Hwang et al. 2001. Bio. Tech. 78: 251-255.

Author Keywords
aboveground biomass, allometric regression equation, Korean pine, natural forest, Pinus koraiensis, plantation, specific leaf area

Key Words Plus
leaf-area, douglas-fir, plantation, nitrogen

Modeling canopy structure and heterogeneity across scales: From crowns to canopy

Bo Song, Jiquan Chen, Paul V. Desanker, David D. Reed

School of Forestry and Wood Products, Michigan Technological University, Houghton, MI 49931, USA

Gay A. Bradshaw

Forestry Science Laboratory, USDA Forest Service, Corvallis, OR 97331, USA

Jerry F. Franklin

College of Forest Resources, University of Washinton, Seatle, WA 98195, USA

Abstract
Canopy studies have been limited in ecological investigations due to problems of canopy accessibility, and the lack of efficient sampling and modeling methods. The primary objective of this study was to develop an efficient modeling approach to describe the 3-dimensional, hierarchical structure of individual crown shells within stands and corresponding canopy patches. Crown shells were modeled based on crown ratio, maximum cardinal radius, vertical position, and shape. Canopies were represented by adding unique crowns to simulated point patterns of trees of known aggregation as measured by Pielou's index of nonrandomness. Canopy patches were delineated at multiple horizontal and vertical scales using the ARC/INFO geographic information system (GIS). The patterns of canopy patches are clearly variable and scale dependent. Canopy patterns become more diverse at broader horizontal scales, and change greatly from the lower to the upper canopies. The modeling approach used in this study has general utility in characterizing 3-dimensional canopies of many types of forests. Song and Chen et al. 1997. For. Ecol. Manage. 96: 217-229.

Author Keywords
crown, canopy, simulation, GIS, scale, spatial pattern

Key Words Plus
forest

The spatial manifestation of forest succession in optical imagery - The potential of multiresolution imagery

Conghe Song

Department of Geography, the University of North Carolina at Chapel Hill, CB# 3220, 203 Saunders Hall, Chapel Hill, NC 27509, USA

Curtis E. Woodcock

Department of Geography, Boston University, 675 Commonwealth Avenue, Boston, MA 02215, USA

Abstract
Forest succession is a fundamental ecological phenomenon, which has significant implications for sustainable ecosystem management as well as biological, biophysical, and biogeochemical processes. Remote sensing is perhaps the only viable option for monitoring changes in forest ecosystems over large areas in a timely and cost efficient manner. This study investigates the spatial manifestation of forest succession in optical imagery through three types of models: a two-component spatial model, a canopy reflectance model (Geometric-Optical and Radiative Transfer, GORT) and a forest ecosystem dynamics model (ZELIG). The latter two models provide inputs to the former one to predict the spatial properties of images as a function of the combined effects of tree size and density, the spectral signatures of scene components and pixel size. An important source of information that is diagnostic of canopy structure has been identified: the spatial properties of multiresolution imagery. The sill of variograms of images of forest stands decrease with regularization, and in particular the rate of decrease is related to the size of trees. For stands with larger trees the sills of variograms decrease more slowly with increasing regularization than for stands with smaller trees. However, the spatial patterns for a scene with multiresolution imagery are also dependent on tree cover. This implies that the use of spatial patterns to estimate tree size will require independent estimates of tree cover as a preliminary step. Concept verification with an Ikonos 1-m panchromatic image for stands at the H.J. Andrews Experimental Forest in the Cascade Range of Oregon indicates the simulated spatial patterns exist in multiresolution imagery. This study demonstrates the potential to map tree size automatically from multiresolution imagery. Song and Woodcock. 2002. Remote Sens. Environ. 82: 271-284.

keywords plus
bidirectional reflectance, coniferous forests, digital images, gap model, canopy, vegetation, landsat, resolution, scale, semivariogram

Sequential application of simplex designs in optimisation and evolutionary operation

W. Spendley, G.R. Hext, F.R. Himsworth

Imperial Chemical Industries Ltd., Biullingham, Co. Durham, England

Abstract

A technique for empirical optimization is presented in which a sequence of experimental designs each in the form of a regular or irregular simplex is used, each simplex having all vertices but one in common with the preceding simplex, and being completed by one new point. Reasons for the choice of design are outline, and a formal procedure given. The performance of the technique in the presence and absence of error is studied and it is shown (a) that in the presence of error the rate of advance is inversely proportional to the error standard deviation, so that replication of observations is not beneficial, and (b) that the “efficiency” of the technique appears to increase in direct proportion to the number of factors investigated. It is also noted that, since the direction of movement from each simplex is dependent solely on the ranking of the observations, the technique may be used even in circumstances when a response cannot be quantitatively assessed. Attention is drawn to the ease with which second –order designs having the minimum number of experimental points may derived from a regular simplex, and a fitting procedures which avoids a direct matrix inversion is suggested. In a brief appendix one or two new rotatable designs derivable from a simplex are noted. Spendley and Hext et al. 1962. Technometrics. 4: 441-461.

Distribution of biomass in an Indiana old-growth forest from 1926 to 1992

Martin A. Spetich, George R. Parker

Department of Forestry and Natural Resources, Purdue University, West Lafayette, Indiana 47907

Abstract
We examined the structural and spatial distribution of woody biomass in relationship to disturbance in an Indiana old-growth deciduous forest over a 66-yr period. Analysis was done on the core 7.92 ha of a 20.6-ha forest in which every tree 10 cm dbh and over has been tagged and mapped since 1926. Five years are compared-1926, 1976, 1981, 1986 and 1992. Dry weight of living biomass for the 7.92-ha area for these 5 vr was 154 Mg/ha, 207 Mg/ha, 220 Mg/ha, 216 Mg/ha and 211 Mg/ha, respectively Biomass of dead trees was 1 Mg ha(-1) yr(-1) from 1977 through 1981; 4 Mg ha(-1) yr(-1) from 1982 through 1986; and 3 Mg ha(-1) yr(-1) from 1987 through 1992. Biomass of trees that died between 1976 and 1992 was greatest for midseral species. Living biomass of dominant early to midseral species is declining while that of late seral species is increasing. In 1926 biomass of trees 10 to 25 cm diam consisted of 14% Quercus spp. and 12% Acer saccharum. By 1992 biomass in this diameter range consisted of 1% Quercus spp. and 43% A. saccharum. Equilibriun parch size was estimated for biomass at each of the five inventory dates to determine ii there was a change. Equilibrium patch size for biomass was estimated to be 0.64 ha during all five inventory dates based on the coefficient of variation (CV) of biomass for 16 different grid cell sizes. Grid cell size refers to the size of adjacent cells in a grid that covered the entire study area. The grid with the smallest cells had cells of 0.01 ha. This grid of 0.01-ha cells was aggregated to 15 additional grid cell sizes, where the largest grid cell size was 1.98 ha. CV for all grid cell sizes was highest in 1926 due to effects of prior grazing. There data indicate an increase in deadwood biomass, a drift in stand composition, recovery from grazing by an increase in small diameter trees and no change in equilibrium patch size over the five inventory dates. Spetich, and Parker. 1998. The American Midland Naturalist. 139: 90-107.

Key Words Plus
coarse woody debris, deciduous forest, oak forest, dynamics, disturbance, regeneration, landscape, patterns, decomposition, replacement

Plot Size Recommendations for Biomass Estimation in a Midwestern Old-Growth Forest

Martin A. Spetich

Southern Research Station, USDA Forest Service, Hot Springs, AR 71902

George R. Parker

Department of Forestry and Natural Resources, Purdue University, West Lafayette, IN 47907.

Abstract

We examine the relationship between disturbance regime and plot size for woody biomass estimation in a midwestern old-growth deciduous forest from 1926 to 1992. Analysis was done on the core 19.6 ac of a 50. 1 ac forest in which every tree 4 in. dbh and greater has been tagged and mapped since 1926. Five windows of time are compared-1926, 1976, 1981, 1986 and 1992. The most efficient plot sizes requiring measurement of the least total area range from 0.02 to 1.6 ac. A small plot size was recommended after grazing disturbance, and a large size was recommended when mortality was high and acceptable error was low. Spetich and Parker. 1992. North J. Appl. For. 15: 165-168.

Height-age trends from an Arkansas seed source study

Peter T. Sprinz

Department of Forest Science, Texas Agricultural Experiment Station, Texas A&M University System, College Station, TX 77843

Mike R. Strub

Biological Sciences Research Division, Weyerhaeuser Company, Tacoma, WA 98477

Cheryl B. Talbert

Southern Forestry Research Division, Weyerhaeuser Co., Hot Springs, AK 71902

Abstract

An analysis of average and dominant height, and height growth indicated the geographic origin of planted loblolly pine significantly affected height development through 29 years in Arkansas. For average and dominant height, sources west of the Mississippi were shorter than eastern sources at all ages; coastal Carolina and Virginia sources were taller than local sources from Arkansas, northern Louisiana, east Texas, and Oklahoma at all ages. The interior sources from northern Mississippi, Alabama, Georgia, and piedmont Carolina were shorter than the coastal sources initially, but became similar by age 25. It was also determined that region rankings for dominant height remained fairly consistent with age, hut the relative magnitudes of difference among regions decreased with age. Region rankings for average and dominant heights were similar, but the relative differences in magnitude were greater for average height than for dominant height. A flexible height-age model suggested that while dominant height maximized at a common level, the shape of development varied to that asymptote with seed source. Lastly, it was found that region differences existed in the shape and level of height growth over time. Eastern origin material showed the greatest height growth initially, but the western origin stands had attained the same height growth rate by age 20. Sprinz and Talbert et al. 1989. For. Sci. 35: 677-691.

Additional Key Words

Pinus taeda, site index, genetic gains

Data management procedures in ecological research

S.G. Stafford, P.B. Alaback, K.L. Waddell, R.L. Slagie

Abstract

Ecological research requires a flexible, organized system for acquiring, documenting, and managing data. Careful documentation of data, best done through close collaboration of the researcher and data manager, is important if all users are to benefit from a centralized database management (DBM) system. A systematic approach to data management and analysis comprises four key steps: comprehension, planning, execution, and evaluation and interpretation. Statistical consulting at the beginning of a project helps scientists plan well-designed, efficient research strategies. Data collection forms should be designed to encourage recorders to enter all essential identifying information, thereby minimizing errors. Consistent, high-quality data verification, through either the visual or double-entry method, allows most data collection errors to be caught and corrected before analysis. Standardized forms help maintain uniformity in data documentation. Documentation and corresponding data files should be linked in a carefully organized relational DBM system in which all information may be easily stored and retrieved. Using commercial software for data management and analysis is usually more cost effective than developing and maintaining customized software. Scarce DBM resources should be invested in data validation, equipment acquisition, system design, and data documentation, not programming. Database maintenance is costly, but critical, to keeping credibility with users. The key to future advances in ecological data management lies in the ability of data managers and scientists to move into a more cooperative, integrative mode through which comprehensive databases are established and more fully used, increasing overall research efficiency.

Prediction of height increment for models of forest growth

Albert R, Stage

Forest Sciences Laboratory, Moscow, Idaho

Abstract

How fast do trees grow in height? The answer requires a great deal of detail about the tree's environment and about the tree itself, or some specific growth measure on the tree that indicates the combined effects of all environmental factors and the tree's individual characteristics. Diameter increment is such a measure. It is much more easily measured than height increment, and responds to the same growth determinants. Diameter increment, however, is much more responsive to effects of stand stocking than is height increment. Past effects of stocking on height growth relative to diameter growth are indicated by the tree's form, as measured by the height/d.b.h. ratio, and by its crown ratio (live crown length/total height). In this report, we seek to develop a prediction equation that relates height increment to concurrent diameter increment, tree height, diameter, crown ratio, and habitat type (Daubenmire and Daubenmire 1968). To use this equation for calculating the height increment component of current volume increment of trees on inventory plots, a computer subroutine is provided in the appendix. For more extended prognoses, the same equation for predicting height increment can be imbedded in a computer model of forest stand development (Stage 1973) For many years Forest Survey in the Intermountain Forest and Range Experiment Station has calculated current volume increment from changes in height computed from diameter increment and diameter. Wright (1961) recommended a similar procedure. Both approaches were derived from a single curve of total height over d.b.h. by determining the differences i n the curve height for current d.b.h. and for d.b.h. plus diameter increment. Hence, these approaches assume that all trees of a particular species follow height/diameter curves that have the same slope for a given d.b.h. Better prediction equations should be possible where data are available from direct measurements of height increment. Prediction equations presented in this report are based on analysis of 1.165 trees felled as part of management planning inventories of the Kaniksu, Coeur d'Alene. St. Joe. and Lewis and Clark National Forests in northern Idaho and western Montana. Stage 1975. USDA For. Serv. Res. Pap. INT-164.

A mathematical approach to polymorphic site-index curves for grand fir

Albert R. Stage

Intermountain Forest and Range Experimental Station, Forest Service, USDA, Moscow, Idaho

Abstract

Polymorphic site index curves may be derived for Abies grandis by fitting to stem analysis data the sigmoid function: H = a exp(-b/A^c). This procedure uses the tree as a phytometer to estimate site index. To arrive at a site quality measure closely corresponding to periodic stand increment, the estimation procedure compensates for the effect of site factors that affect the tree only during the seedling and sapling stages of development. Site index is here defined by the annual height increment on a tree 55 feet tall, but expressed as tree height at an index of 50 years. The number of annual rings in 1.5-inch radius from the pith at breast height is used to select the appropriate set of height/age curves from the polymorphic series. Such curves constructed from stem analyses must be applied to a uniform number of the tallest trees per acre at all ages. This number is tentatively set at the rate of 25 trees per acre. Stage 1963. For. Sci. 9: 167-180.

Dominant-height and site-index equations for Ponderosa pine in east-central Arizona

W. F. Stansfield, J. P. Mctacue

Northern Arizona University, Flagstaff, A Z 86011, U.S.A

R. Lacapa

Division of Forestry, Fort Apache Indian Agency, Bureau of Indian Affairs, Whiteriver, A Z 85941, U.S.A

Abstract
A dominant-height equation for ponderosa pine (Pinus ponderosa Laws.) was constructed utilizing a parameter prediction method and a model by J.E. King. The site-index equation was obtained by inverting the dominant-height equation. A method was examined for indirectly obtaining parameter prediction equations. It proved superior to two direct parameter prediction approaches. Other site-quality variables, such as habitat type groups and elevation, were included in the dominant-height and site-index equations and were successful in refining predictive capability. Stansfield and McTague et al. 1991. Can. J. For. Res. 21: 606-611.

Key Words Plus
forest

Combining multiple classifiers: An application using spatial and remotely sensed information for land cover type mapping

Brian M. Steele

Department of Mathematical Sciences, University of Montana, Missoula, Montana, MT 59812

Abstract
This article discusses two new methods for increasing the accuracy of classifiers used land cover mapping. The first method, called the product rule, is a simple and general method of combining two or more classification rules as a single rule. Stacked regression methods of combining classification rules are discussed and compared to the product rule. The second method of increasing classifier accuracy is a simple nonparametric classifier that uses spatial information for classification. Two data sets used for land cover mapping of Landsat TM scenes from Idaho and Montana illustrate the new methods. For these examples, the product rule compared favorably to the more complex stacked regression methods. The spatial classifier produced substantial increases in estimated accuracy when combined with one or more additional classifiers that used remotely sensed variables for classification. These results suggest that the product rule may produce increases in map accuracy with little additional expense or effort. The spatial classifier may be useful for increasing accuracy when patterns exist in the spatial distribution of land cover. Steel 2000. Remote Sens. Environ. 74: 545-556.

keywords plus
contextual classification, sensing data, images

Practical implications of design-based sampling inference for thematic map accuracy assessment

Stephen V. Stehman

State University of New York, College of Environmental Science and Forestry, 320 Bray Hall, Syracuse, NY 13210

Abstract
Sampling inference is the process of generalizing from sample data to make statements or draw conclusions about a population. Design-based inference is the inferential framework commonly invoked when sampling techniques are used in thematic map accuracy assessment. The conceptual basis of design-based inference is described, followed by discussion of practical implications of design-based inference, including (1) the population to which the inferences apply, (2) estimation formulas and their justification, (3) interpretation of accuracy measures, (4) representation of variability, (5) effect of spatial correlation, and (6) role of probability sampling. Design-based inference is contrasted with model-based inference, another inferential framework frequently invoked in statistics. Stehman 2000. Remote Sens. Environ. 72: 35-45.

keywords plus
remotely-sensed data, classification accuracy, sensing data, cover map, probabilities, uncertainty, coefficient, perspective, errors, basin

Design and analysis for thematic map accuracy assessment: fundamental principles

Stephen V. Stehman

SUNY College of Environmental Science and Forestry, Syracuse, New York

Raymond L. Czaplewski

USDA Forest Service, Rocky Mountain Research Station, Fort Collins, Colorado

Abstract
Before being used in scientific investigations and policy decisions, thematic maps constructed from remotely sensed data should be subjected to a statistically rigorous accuracy assessment. The three basic components of an accuracy assessment are: 1) the sampling design used to select the reference sample; 2) the response design used to obtain the reference land-cover classification for each sampling unit; and 3) the estimation and analysis procedures. We discuss options available for each of these components. A statistically rigorous assessment requires both a probability sampling design and statistically consistent estimators of accuracy parameters, along with a response design determined in accordance with features of the mapping and classification process such as the land-cover classification scheme, minimum mapping unit, and spatial scale of the mapping. Stehman, and Czaplewski. 1998. Remote Sens. Environ. 64: 331-344.

Key Words Plus
remotely-sensed data, rural-urban fringe, classification accuracy, satellite imagery, land-use, improve, forest, error, resources, size

An overview of spatial sampling procedures and experimental design of spatial studies for ecosystem comparisons

Alfred Stein

Wageningen University, Mathematical and Statistical Methods Group, Dreijenlaan 4, 6703 HA Wageningen, The Netherlands

Christien Ettema

Wageningen University, Sub-Department of Soil Quality, Dreijenplein 10, 6703 HB Wageningen, The Netherlands

Abstract
Comparison of ecosystems and land use studies often require the use of non-classical statistics. This paper describes modem ways of approaching optimal sampling for ecological and environmental purposes. The first part of the paper deals with a description of different sampling procedures. A distinction is made between sampling surveys, optimal grid spacing and adaptive sampling. The second part of the manuscript uses a simulated example to illustrate the different sampling procedures. The third part contains an actual field study where various constraints were met that had to be incorporated in the lay out of the sampling scheme. Optimal statistical sampling procedures can be implemented and utilized to collect maximum information from limited resources. Stein and Ettema. 2002. Agr. Ecosyst. Environ. 1939: 1-17.

Author Keywords
optimal sampling, adaptive sampling, bio-diversity, experimental design, soil biota, spatial simulations

Key Words Plus
optimal interpolation, soil properties, optimization, strategy

Modeling space and time dependence in environmental studiesModeling space and time dependence in environmental studies

Alfred Stein

ITC, PO Box 6, 7500 AA Enschede, The Netherlands

Dept. of Environmental Sciences, Agricultural University, PO Box 37, 6700 AA Wageningen, The Netherlands

Geert Sterk

Dept. of Environmental Sciences, Agricultural University, PO Box 37, 6700 AA Wageningen, The Netherlands

Abstract

This paper analyzes the use of stratification in the modeling of dependence for regionalized variables in space and time. Stratification enables optimal use of available information, and it is used when deterministic, small-scale information is combined with large-scale random variation: on the one hand, the problem of having too few data to reliably estimate statistical parameters may be solved by stratification; on the other hand some variation may be so obvious that it serves as a prerequisite for obtaining any form of stationarity. The paper is illustrated with four case-studies. The first study analyzes optimization of an extant groundwater monitoring network within a catchment of the river Rhine. Stratification in space enabled 5 cm precision of predicted levels at critical locations, something that could not be reached without stratification. The second study analyzed variability in space and time of wind erosion in a field. Erosion by four storms was measured with 21 devices. Use of temporal stratification to overcome lack of spatial data resulted, in one instance, in a decrease of 17 percent in mean absolute error. The third study addressed changes in the soil solution influenced by environmental deposition over a period of six years. Stratification was made in space to enable analysis in the time domain and to obtain spatio-temporal maps. The fourth study used a general space-time model to evaluate the development of root-rot in soybean, caused by naturally occurring pathogens. Predictions were made at unvisited times to estimate the probability of occurrence of root rot. We conclude that use of stratification enhances the utility of space-time statistics for the solution of various environmental and agricultural problems. Stein and Sterk 1999. JAG. 1: 109-121.

Key Words

stratification, space-time statistics, wind erosion, groundwater level, soil solution, phytopathology

Spatial variability of soil properties at different scales within three terraces of the Henares River (Spain)

A. Saldana, A. Stein, J.A. Zinck

ITC International Institute for Aerospace Survey and Earth Sciences, P.O. Box 6, 7500 AA Enschede,

Netherlands

Six-year growth of Douglas-fir saplings after manual or herbicide release from coastal shrub competition

William I. Stein

Forestry Sciences Laboratory, 3200 SW Jefferson Way, Corvallis, OR 97331

Author Keywords
chronosequence, spatial variability, variogram, soil homogenisation, spatial sampling

Key Words Plus
geostatistics, delineations, information, landscapes

Growth of white spruce following release from trembling aspen

G.A. Steneker

Department of Forestry and Rural Development, Forest Research Laboratory, Winnipeg, Manitoba

Abstract

Experimental release cuttings to favor white spruce in 15- to 60-year-old white spruce- trembling aspen stands have shown that diameter increment and in certain instances height increment of the spruce was doubled; merchantable volume production per ' acre was, on the average, increased by about 60 percent. These findings are of special interesting the Prairie Provinces, where the spruce-aspen cover type forms the principal source of white spruce. Steneker 1967. For. Branch Dep. Pub. Gen. Tech. Rep. 1183.

Spatial variation of nitrate-N and related soil properties at the plot-scale

R. Stenger

Institute of Soil Ecology, GSF-National Research Center for Environment and Health, D-85764 Neuherberg, Germany

Lincoln Environmental, Private Bag 3062, Hamilton, New Zealand

E. Priesack

Institute of Soil Ecology, GSF-National Research Center for Environment and Health, D-85764 Neuherberg, Germany

F. Beese

Institute of Soil Science and Forest Nutrition, University of Gottingen, Germany

Abstract
Neglecting the spatial variation in soil nutrient status may result in unused yield potential and in environmental damage. Site-specific management has been suggested to reduce inappropriate fertilization that can adversely affect soil, ground and surface water. Decision criteria for determining variable-rate nitrogen fertilization are, however, lacking. This paper analyses the spatial variation of nitrate nitrogen (NO3-N) and soil properties related to the N cycle at the plot-scale, Three 50 X 50 in plots were sampled in nested sampling designs of varying complexities. Classical statistics revealed a characteristic ranking in the variability of soil properties. Geostatistical analysis of the NO3-N data from two plots showed that the small-scale variation found in one small subgrid was not typical for the small-scale variation in the entire plot, indicating bias in the sampling design. A trend component was found in the NO3-N data and, consequently, the minimal requirement for the regionalized variable theory was not fulfilled. Problems due to design were overcome with a more complex nested sampling at the third plot. However, the spherical model fitted to the NO3-N data of the first year explained only 21% of the total variance, whereas a pure nugget effect was observed in the second year. The water content data also showed a low structural variance, which was different in the two years, In contrast, two thirds of the variance of total carbon (C) and total nitrogen (N) could be explained by the fitted models. Seasonal variations, such as varying duration of snow cover, and extrinsic management effects, such as growing of a cover crop, may have contributed to the observed differences in variability between the years. Due to the low proportion of structural variance and the observation that spatial distribution was not stable with time, geostatistical analysis of NO3-N and water contents data added only little information to classical statistical analysis. However, geostatistical analysis of total C and N contents provided a useful means to calculate spatial distribution patterns of these properties. Stenger and Priesack et al. 2002. Geoderma. 105: 259-275.

Author Keywords
site-specific management, spatial variation, temporal variation, heterogeneity

Key Words Plus
temporal persistence, field-scale, variability, patterns, nitrogen, mineralization, availability, dependence, management

Adapting an individual tree growth model for Norway spruce (Picea abies L. Karst.) in pure and mixed species stands

Hubetr Sterba, Astrid Blab

Institute of Forest Growth Research, University of Agricultural Sciences, Peter Jordanstrabe 82, A-1190 Wien, Vienna, Austria

Klaus Katzensteiner

Institute of Forest Ecology, University of Agricultural Sciences, Peter Joudanstrabe 82,A-1190, Wien, Vienna, Austria

Author Keywords
individual tree growth model, mixed species stands, Picea abies L. Karst., Fagus sylvatica L.

KeyWords Plus
forest stands, beech

Crown efficiency in a loblolly pine (Pinus taeda) spacing experiment

Hubert Sterba

Institut fur Waldwachstumsforschung, Universitat fur Bodcnkultur Wien Peter JrdanstranBe 82, A-1190 Vienna, Austria

Ralph L. Amateis

Department of Forestry, Virginia Tech. Blacksburg, VA 24061-0324, USA

Abstract
Crown efficiency was first defined by Assmann (1961. Waldertragskunde. BLV, Munchen) as individual tree volume increment per unit of crown projection area. He hypothesized that within a given crown class, smaller crowns are more efficient because their ratio between crown surface and horizontal crown projection is higher. Data from a loblolly pine (Pinus taeda L.) spacing experiment were used to test if this hypothesis also holds in young loblolly pine stands and, if so, to determine if it explains the increment differences between spacings in the spacing experiment. Using individual tree height relative to plot dominant height to describe crown class, within-plot regression showed that crown efficiency decreased with crown size for trees below dominant height. This relationship was much less pronounced than indicated from Assmann's examples, although the crown surface to crown projection ratio behaved in the same way as Assmann had hypothesized. Crown efficiency as well as the crown surface to crown projection area ratio decreased with increasing density. Basal area increment per hectare increased until total crown closure approached 130% and then stayed constant. This major impact of total crown coverage brings into question the usefulness of crown efficiency as an indicator for unit area growth. Sterba and Amateis. 1998. Can. J. For. Res. 28: 1344-1351.

Key Words Plus
growth efficiency

Sampling In Forest Experiments

W.G. Stevenson, P.S. Savill

Biometrics Division and Forest Service Department o f Agriculture Northern Ireland

Abstract

In a computer simulation exercise, samples of trees of various sizes were drawn from assessment plots of 36 trees in two Sitka spruce experiments. Experimental means from the different sample sizes never deviated far from the statistics of the analyses of the complete assessment plots variances decreasing uniformly with increasing sample size. The observed residual mean squares corresponded very closely with the expected values. It is concluded that little extra information is gained by measuring more than 20 trees in any treatment plot and that as few as 12 will give acceptable results. Stevenson and Savill. 1976. Commonw. For. Rev. 55. 4: 331-334.

Improvement cut accelerates white pine sawlog growth

W.M. Stiell

Petawawa National Forestry Institute, Chalk River, Ontario

Abstract

Middle-aged eastern white pine (Pinus strobes L.) were released in 1971 by removal of overstory hard-woods in a commercial cut. Ten-year growth of the pine indicated that treatment had been effective in promoting development of sawlog sizes. Compared with untreated stands, the margin of growth of ths material was up to 30 m³/ha for middle-density under-storeys (12 m²/ha of basal area), and would reach 71 m³/ha over 20 years according to growth projections. Stiell 1984. For. Chron. 60: 3-9.

How uniformity of tree distribution affects stand growth

W.M. Stiell

Department of Fisheries and Environment, Canadian Forestry Service, Petawawa Forest Experiment Station, Chalk River, Ontario

Abstract

Part of a 13-year-old plantation of red pine (Pinus resinosa Ait.) was thinned to leave stems uniformly distributed throughout the stand and another part was thinned to leave the same number of stems but in equally spaced 4-tree clumps. After 10 years the clumped trees had smaller live crowns (owing to more intense shape on the inside of the clumps reflected in lower growth in d.b.h. and basal area per hectare, and higher form class. No significant difference between treatments was found in volume per hectare. But it was concluded that 5-treeor larger clumps would result in reduced wood production. Stiell 1978. For. Chron. 54: 156-158.

Yield of unthinned red pine plantations at the Petawawa forest experiment station

W.M. Stiell

Forest Management Institute Ottawa, Ontario

A.B. Berry

Petawawa Forest Experiment Station Chalk River, Ontario

Abstract

Yield tab3es for high-survival, unthinned plantations of red pine were prepared to age 50 years from planting, by 5-year age classes, for eight planted spacing and five site index classes. Each table presents the numbers of trees and the basal area per acre, the mean d.b.h., and the total, merchantable cubic, and board foot volumes per acre. Site index curves and diameter distribution data are also presented. Stiell and Berry. 1973. Dept. Environ. Can. For. Serv. Publ. NO. 1320.

Some competitive relations in a red pine plantation

W.M. Stiell

Department of Fisheries and Forestry, Canadian Forestry Service, Petawawa Forest Experiment Station, Chlak River, Ontario

Abstract

Thirteen-year-old red pine (Pinus resinosa Ait.), planted at 5’X5’ and 7’X7’ spacings on sandy, old-field sites, were afforded various types of release from competition by bending over or felling their immediate neighbors, or by severing (but leaving upright) the tops of adjacent trees. Others were subjected to increased aerial competition by inserting severed tops in an upright position around them. The intention was to differentiate competitive effects of crowns from those of roots, and response to treatment was assessed by 2-year stem volume growth. The results could not be fully explained in terms of the treatments applied until hydraulic excavation revealed that root systems were overlapping and very widely dispersed, extending over many times the area occupied by the respective crowns. It is concluded that while above-ground competitions concentrated within the areas occupied by individual crowns, root competition is diffuse and its origin unpredictable for a given tree. Some of the implications for practical silviculture are presented. Stiell 1970. Dept. Fish. and For., Can. For. Serv. Publ. NO. 1275.

Red pine crown development in relation to spacing

W.M. Stiell

Department of Forestry of Canada, Petawawa Forest Experiment Station Chalk River, Ontario

Abstract

Crown and stem measurements were taken eriodically on selected trees in plantations of red pine ( Pinus Resinosa Ait.) over the period 17 ot 37 years from seed. Spacings were 5X5, 7X7, 10X10, and 14X14 feet; some of these represented the original planting interval and others had been achieved by mechanical thinning at the start of the experiment. Little mortality occurred during the 20 years, and individual crown size tended to stabilized, with smaller crowns occurring with closer spacings. Periodic volume growth per acre was strongly correlated with average foliage weight for the period. Thinning was followed by an immediate but temporary increase in foliage efficiency, and wood production per kilo oven dry foliage ranged from 0.023 to 0.038 cubic feet. Crown length, width, and weight, once stabilized, did not show good correlation with parameters of individual stem growth which are cumulative. Height to crown continued to increase and was strongly correlated with total volume per acre and with the ratio average height/(average spacing)^1/3. Breast-height growth of sample tress appeared to be a function of height to crown and foliage weight. Stiell 1966. Can. Dep. For., Publ. 1145.

The process and progress of economics

George J. Stigler

University of Chicago

Abstract

The lecture focuses on the reasons that new ideas are accepted or rejected by a science. A distinction is drawn between prescientific and scientific stages of a discipline. The diverse fates of new ideas are illustrated by a variety of episodes in the history of economics, including the economics of information and the theory of economic regulation. Stigler 1983. J. Polit. Econ. 91: 529-545.

Plantation thinning systems in the southern United States

Bryce J. Stokes

USDA Forest Service, Auburn, Alabama, USA

William F. Watson

School of Forest Resources, Mississippi State University, Mississippi State, Mississippi, USA

Abstract

This paper reviews southern pine management and thinning practices, describes three harvesting systems for thinning, presents production and cost estimates, and utilization rates. The costs and product recoveries were developed from published sources using a spreadsheet analysis. Systems included tree-length, flail/chip, and cut-to-length. The estimated total harvesting, transport, and woodyard cost per m³ of pulpable fiber at the digester was US$24.15 for tree-length and US$19.84 for flail/chip. The same costs for cut-to-length was US$27.66, 27.87, and 29.15 per m³ for chainsaw, feller-buncher/processor, and harvester systems, respectively when producing 2.3-m boltwood. When processing the trees into 5.3-m bolts, the cost for the harvester system was US$29.05. Fiber recovery to the digester was approximately 55 percent of standing biomass for all the systems. Stokes and Watson. 1996.

Site quality estimation using height and diameter

Benjamin B. Stout

School of Forestry, University of Montana, Missoula MT 59812

Durland L. Shumway

Concrete, WA 98237

Abstract

A method for estimating site quality using height and diameter of dominant and codominant trees independent of tree age is examined. Easter hardwoods are the test species. Two parameters define the H-D site curves: a site parameter and a species parameter. The estimated site parameters increase with increasing site index and are significantly different between site classes. The estimated species parameters are species specific and are not significantly different between site classes. Curves for estimating site class are presented. Some species reach a higher proportion o f total height at a given diameter. The effects of these apparent height growth patterns are discussed from a silvicultural viewpoint. Stout and Shumway. 1982. For. Sci. 28: 639-645.

Additional Key Words

Site index, hardwoods

Satellite-derived vegetation index and cover type maps for estimating carbon dioxide flux arctic tundra regions

Douglas Stow, Allen Hope, William Boynton, Stuart Phinn

Department of Geography, San Diego State University, San Diego, CA 92182 4493, USA

Donald Walker, Nancy Auerbach

Institute of Arctic and Alpine Research, University of Colorado, Boulder, CO 80309, USA

Author Keywords
arctic tundra, remote sensing, carbon dioxide, regional scaling, satellite imagery, spectral vegetation indices

Key Words Plus
reflectance, ecosystems, co2, exchange

Comparison of diameter growth and crown competition factor in loblolly pine plantations

Mike R. Strub

School of Forest Resources, Pennsylvania State University, University Park, PA 16802

Richard B. Vasey, Harold E. Burkhart

Division of Forestry and Wildlife Resources, Virginia Polytechnic Institute and State University (VPI & SU) Blacksburg, VA 24061

Abstract

Growth in diameter at breast height (dbh) in loblolly pine plantations was compared with that of open-grown trees. Data were collected in 240 plantations and from 81 open-grown pines. The age at which average diameter growth in plantations diverges from that of open-grown trees was estimated for different site index and stand density combinations and compared to the age at which the crown competition factor (CCF) is predicted to reach 100. Average diameter growth in stands consistently differed from open-grown tree diameter growth- one year after the CCF reaches 100. This result is potentially useful for forest growth simulation. Strub and Vasey et al. 1975. For. Sci. 21: 427-431.

Additional key words

Pinus taeda, competition

Height-age trends from an Arkansas seed source study

Peter T. Sprinz

Department of Forest Science, Texas Agricultural Experiment Station, Texas A&M University System, College Station, TX 77843

Cheryl B. Talbert

Biological Sciences research Division, Weyerhaeuser Company, Tacoma, WA, 98477

Mike R. Strub

Southern Forestry Research Division, Weyerhaeuser Co., Hot Springs, Ak 71902

Abstract

Additional Key Words

Pinus taeda, site index, genetic gains.

Lodgepole pine regeneration in an old, self-perpetuating forest in South central Oregon

John D. Stuart

College of Natural Resources, Humboldt State University, Arcata, CA 95521, U.S.A.

James K. Agee

National Park Service Cooperative Park Studies Unit, College of Forest Resources, University of Washington, Seattle, WA 98195, U.S.A.

Robert I. Gara

College of Forest Resources, University of Washington, Seattle, WA 98195, U.S.A

Abstract

Historic regeneration patterns and regeneration requirements were investigated in an old, self-regenerating lodgepole pine (Pinus contorta Dougl. ssp. murrayana (Balf.) Critchfield) forest in south central Oregon. The forest was multiaged, with episodic regeneration pulses being correlated with mountain pine beetle (Dendroctonus ponderosae Hopk.) outbreaks or fire. The magnitude of a regeneration pulse was a function of disturbance intensity. Tree ring indices show growth declines prior to mountain pine beetle outbreaks. Radial tree growth improved following disturbance. Differences in stand structure among climax lodgepole pine stands in the Rocky Mountains, Sierra Nevada, and south central Oregon were related to disturbance type, frequency, and intensity. Successful lodgepole pine reproduction was limited by soil moisture and partly by microclimate. Shading did not inhibit seedling establishment, but rather provided relief from excessive evapotranspiration, heat, and frost. Stuart and Agee et al. 1989. Can. J. For. Res. 19: 1096-1104.

Regression Problems from Repeated Measurements

Alfred D. Sullivan

Department of Forestry, Mississippi State University, Mississippi State, 39762

Marion R. Reynolds, Jr.

Professor of Statistics and Forestry, Virginia Polytechnic Institute and State University, Blacksburg, 24061

Abstract

Some consequences of using ordinary least squares estimation procedures with remeasurement data are explained. Of particular concern is the bias of variance estimates and alteration of the confidence coefficient for confidence intervals which result from correlations in the data. Results are presented allowing consideration of the problem's severity for data with two repeated measurements. Finally, alternative estimation procedures including generalized least squares and maximum likelihood are discussed. Sullivan and Reynolds Jr. 1976. For. Sci. 22: 382-385.

Additional key rods

Permanent plot analysis, correlated data