|David Walker||Data needs and availability for developing and testing national scale biomass estimators||Tuesday||3c||Sequoyah||Development and implementation of new volume/biomass/carbon models for the FIA program||Felled-tree records were compiled from legacy sources and new field campaigns in order to build a database suitable for developing and testing models for use in FIA national volume, biomass and carbon inventories. Some of the issues associated with the complex data set will be addressed, including the use of both US and Canadian data, dealing with missing observations within trees, and existing gaps in tree species, size, and specific geographic locations where no data are yet available.||A two-track approach involving legacy data compilation and new data collection was pursued to provide sufficient data for developing and testing models to use in FIAs national-scale biomass estimation. Legacy data sources included agency, university, and industry studies from the US, as well as government studies from both provincial and national agencies in Canada. The main data type involved stem taper and volume measurements, mainly from felled-tree studies conducted over the past 120 years in the US and Canada. Weight and biomass data including wood properties such as specific gravity comprised the second largest type of data, almost exclusively from felled-tree studies conducted in North America over the past 60 years. Major gaps in legacy data tree species, size, and geographic location were addressed by conducting field campaigns at strategic locations in the US over a period of about eight years since 2011. In total, data from nearly a quarter-million trees were compiled, about one-tenth of which include biomass measurements. Despite existing needs to be addressed in ongoing model development, the database is among the largest known collections of felled-tree records ever compiled for national-scale model development.|
|David Affleck||Alternative Modeling Strategies for Estimating Tree Biomass Across a Nationwide System of Inventory Plots||Tuesday||3c||Sequoyah||Development and implementation of new volume/biomass/carbon models for the FIA program||We review the biological, practical, and statistical and limitations of alternative model systems for estimating volume and component biomass at the tree-level over the US Forest Inventory & Analysis (FIA) plot network. Developed under FIAs National Biomass Estimators Project include systems that model|
1) total tree mass directly or indirectly;
2) mass as a function of volume, or independently of volume; and
3) trans-species or individual species trends.
|We review the advantages and limitations of alternative model systems for estimating tree volume and biomass over the Forest Inventory & Analysis (FIA) plot network. Systems proposed under FIAs National Biomass Estimators Project have been structured according to biological and dimensional principles. Yet user needs and FIA protocols have also figured prominently in development and evaluation. Chief among the former are demands for being data-driven, for additivity among biomass components, and for smoothness across administrative boundaries, as well as the desirability of compatibility between stem volume and mass estimates. Considerations related to inventory protocols include retrospective applicability and the economics associated with collection of alternative variables. Similarly, the form, quantity, and distribution of tree data, as well as inherent data dependence structures, have motivated alternative statistical specifications. We thus describe the biological, practical, and statistical strengths of systems that model
1) total tree mass directly or indirectly;
2) mass as a function of, or independently of, volume; and
3) trans-species or individual species trends.
Instances of these are explored further in this session.
|Philip Radtke||Evaluating Modeling systems for National-Scale Biomass Estimators: A Scorecard Approach with Preliminary Results||Tuesday||3c||Sequoyah||Development and implementation of new volume/biomass/carbon models for the FIA program||Modeling approaches for live tree aboveground biomass (AGB) estimators will be evaluated for desirable modeling properties: a) component additivity; b) compatibility of multiple attribute predictions; c) greatest accuracy in predicting AGB for major species; d) performance of models used to predict for minor species; e) well-behaved prediction patterns in models when used to extrapolate to very large trees; and f) effectiveness of incorporating stand or site-level predictors.||Numerous approaches will be evaluated to ensure desirable modeling properties in live tree aboveground biomass (AGB) estimators for national forest and carbon inventories, including: a) component additivity for stem and branch wood or bark, branches, and foliage; b) compatibility of volume, taper, biomass, and specific gravity predictions; c) accuracy in predicting AGB for species well-represented in model fitting data sets and for those lacking data; d) performance of models used to predict for species lacking adequate data; e) well-behaved prediction patterns in models when used to extrapolate beyond the range of observed data; and f) the effectiveness of incorporating stand and site-level predictors to reduce prediction uncertainty.
A scorecard approach has been developed for delivering concise yet complete information to aid in the model evaluation process. The scorecard approach enumerates categorically which properties proposed modeling systems are designed to achieve and provides a means for reporting quantitative measures of models performance. Examination of both the intended design and measured performance of proposed models will lead to a transparent modeling solution with highly favorable predictive abilities.
|David MacFarlane||Functional, species-specific or hybrid groups for new tree models for FIA plots?||Tuesday||3c||Sequoyah||Development and implementation of new volume/biomass/carbon models for the FIA program||This talk describes ideas, data and models are presented from several studies to examine alternative, tree-functional approaches to species-specific models, along with hybrid approaches that use both species and functional types to better characterize tree to tree variation, both within and between species, across the large spatial domain of FIA.||The Forest Inventory and Analysis (FIA) program of the USDA Forest Service is compiling continent-wide data to create new mass and volume models to be applied to trees on FIA inventory plots. FIA will need to consider how these models should be constructed to capture variation in tree attributes across the vast array of tree species, forest ecosystems, climatic regions and forest disturbance regimes that comprise the scope of FIAs inventory. In practical terms, this means allowing model coefficients and model forms to vary and, traditionally, this would mean allowing model coefficients to vary by species or groups of species. Here, ideas, data and models are presented from several studies to examine alternative, tree-functional approaches to species-specific models, along with hybrid approaches that use both species and functional types to better characterize tree to tree variation, both within and between species, across the large spatial domain of FIA. A new species form type volume modeling approach, developed for the Michigan state forest inventory system, is presented, as case study of the potential utility of a species-functional group approach for both volume and mass estimation for trees on FIA plots.|
|Krishna Poudel||Approaches to Estimate Individual Tree Aboveground Biomass||Tuesday||3c||Sequoyah||Development and implementation of new volume/biomass/carbon models for the FIA program||For the past six years, the FIA along with university partners across the nation have worked together to compile long-term legacy data as well as to collect new set of destructively sampled data to develop new models. This unified national-scale dataset was used to test different approaches to estimate total aboveground live tree biomass as well as the biomass of aboveground components.||Forest ecosystems contribute substantially to global climate change mitigation by sequestering and storing carbon. Forest carbon inventories are obtained by using tree and area measurements along with biomass equations. Therefore, large-scale forest biomass estimation is important to assess the role of forestry sector in mitigating climate change impacts. Methods to estimate total aboveground biomass as well biomass of different tree components (stem, bark, branch, and foliage) were developed using large dataset compiled from studies over the years. Missing components in the dataset were imputed using species-specific or combined-species Dirichlet imputation. Alternative model formulations and fitting techniques for obtaining biomass estimates were tested. Specifically, the independent estimation of biomass is compared with the biomass estimates derived from volume estimates that are obtained using volume equations fitted in this study. Error produced by the methods developed in this study are compared with the error produced by the Component Ratio Method currently being used for official U.S. forest carbon inventories.|
|Aaron Weiskittel||National Scale Biomass Estimator (NSBE) Project: Next steps, implications, and future timeline||Tuesday||3c||Sequoyah||Development and implementation of new volume/biomass/carbon models for the FIA program||The National Scale Biomass Estimator (NSBE) has been a joint collaborative project between several universities and the US Forest Service Forest Inventory & Analysis (FIA) program. The overall goal is to refine FIAs approach to estimating aboveground tree biomass and carbon for the primary species in the US using existing and newly collected data. Current efforts have shifted to evaluating alternative have shifted to evaluating alternative modeling approaches based on this available data.||The National Scale Biomass Estimator (NSBE) has been a joint collaborative project between several universities and the US Forest Service Forest Inventory & Analysis (FIA) program. The overall goal is to refine FIAs approach to estimating aboveground tree biomass and carbon for the primary species in the US using existing and newly collected data. Acquiring and strategically collecting biomass data was an initial focus of the project, while current efforts have shifted to evaluating alternative modeling approaches based on this available data. A current target for the project is to deliver a revised methodology to FIA for estimating tree biomass in 2020. Implementation and testing of this revised methodology by FIA will then occur with particular focus on communicating the revisions to FIAs broad array of key stakeholders. However, preliminary results indicate rather significant shifts in total biomass for certain species and regions, which have important implications for US biomass and carbon estimates. This presentation will review past accomplishments, current plans for next steps in the coming year, and the broader implications of revising FIAs approach to estimating biomass nationally.|