Source: MISSISSIPPI STATE UNIV submitted to NRP
ENHANCING FOREST GROWTH AND YIELD MODELS FOR FOREST MANAGEMENT AND INVENTORY IN MISSISSIPPI AND THE SOUTH
Sponsoring Institution
National Institute of Food and Agriculture
Project Status
COMPLETE
Funding Source
Reporting Frequency
Annual
Accession No.
1005335
Grant No.
(N/A)
Cumulative Award Amt.
(N/A)
Proposal No.
(N/A)
Multistate No.
(N/A)
Project Start Date
Jan 12, 2015
Project End Date
Feb 16, 2018
Grant Year
(N/A)
Program Code
[(N/A)]- (N/A)
Recipient Organization
MISSISSIPPI STATE UNIV
(N/A)
MISSISSIPPI STATE,MS 39762
Performing Department
Dept of Forestry
Non Technical Summary
Forest growth and yield models provide estimates of forest stand characteristics at a future point in time and hence information needed to make management and policy decisions about a given forest resource. Growth and yield models are dynamic systems. As more data and knowledge from research become available, models are enhanced to improve their prediction accuracy, provide alternative modeling frameworks that address different user needs, and enable them to make predictions under different management and biotic and abiotic environmental conditions. Models for loblolly, longleaf, and slash pinedeveloped for Mississippi and the mid Gulf region were put together in the 1980s and early 1990s. Since then, significant developments in modeling techniques have occurred and more data has been collected for some of the species. In addition, almost all slash and loblolly pine stands are currentlyestablished using advanced generation genetically improved planting stock.With the significant amount of new knowledge, there is need for model revision. There is also a need for individual tree models of these species to provide a higher resolution modeling alternative. Research effort on modeling of hardwood stands in Mississippi and the mid Gulf region has been small and hence models for hardwoods are still not well developed or are lacking.Objectives: Theobjectives of the researchare to: 1)develop new stand level diameter distribution models for plantation loblolly and slash pine and even-aged natural longleaf pine, 2) develop individual tree models for plantation loblolly and slash pine and even-aged natural longleaf pine, 3)establish new growth and yield plots in managed natural stands of red oak-sweetgum bottomland hardwoods and inLower Mississippi Alluvial Valleyhardwood plantations, 4) propose new ways of accounting for genetic improvement effects in loblolly pine growth and yield models and incorporate some of the approaches in the models developed in Objectives 1 and 2,5)propose the best statistical modeling approaches and model forms for the forest types in Mississippi and the mid Gulf region, and 6) develop new Windows® computer graphic and web-based user interface for the models developed in Objectives 1 and 2.Approach: Existing data and data that will be collected from ongoing and yet to be established studies will be used in modeling activities for developing whole stand diameter distribution models, individual tree models, genetic improvement response functions, and thinning response functions for loblolly, slash, and longleaf pine and hardwood stands in Mississippi and the mid Gulf region. Various regression techniques will be used and techniques of accounting for hierarchical and repeated measures nature of the data will be applied where necessary. The effect of soil and physiographic factors will be investigated. Alternative model formulations from the literature and those that might be developed during this research will be tested and those with the lowest prediction bias and variance and exhibit logical behavior will be adopted. Performance of the model system will be evaluated using cross validation techniques and sensitivity and uncertainty analysis techniques.The developed models will be assembled into computer instalable simulators using thesoftware development tools inMicrosoft Visual Studio® software.
Animal Health Component
50%
Research Effort Categories
Basic
50%
Applied
50%
Developmental
(N/A)
Classification

Knowledge Area (KA)Subject of Investigation (SOI)Field of Science (FOS)Percent
1230611209070%
1230621209030%
Goals / Objectives
Major Goal of the Project:The overal goal of the project is to provide forest landowners and scientists ofMississippi and mid-Gulf region of southern United Stateswith upto date forest growth and yieldmodel simulatorsthat areunbiased and precise, address the different forest types in the region, and providealternative modeling resolutions.The specific objectives are:Objective 1: Develop new stand level diameter distribution models for plantation loblolly and slash pine and even-aged natural longleaf pineObjective 2: Develop individual tree models for plantation loblolly and slash pine and even-aged natural longleaf pineObjective 3: Establish new growth and yield plots in managed natural stands of red oak-sweetgum bottomland hardwoods and in Lower Mississippi Alluvial Valleyhardwood plantationsObjective 4: Propose new ways of accounting for genetic improvement effects in loblolly pine growth and yield models and incorporate some of the approaches in the models developed in Objectives 1 and 2Objective 5: Propose the best statistical modeling approaches and model forms for the forest types in Mississippi and the mid Gulf region.Objective 6: Develop new Windows® computer graphic and web-based user interface for the models developed in Objectives 1 and 2
Project Methods
Objective 1: Stand level diameter distribution models for plantation loblolly and slash pine and even-aged natural longleaf pineDataData that were used to develop the CLob and NLongleaf models are currently available. These will, respectively, be used to develop the new plantation loblolly and the natural longleaf pine models. For the longleaf pine model, data from an additional 134 plots are available through a data use agreement with the US Forest Service and these will be included in the development of this model. Data that were used to develop the CSlash model will be used for the new slash pine model. These data are currently being sought from the US Forest Service.AnalysisIndividual model components will be fitted using various regression techniques. Techniques of accounting for hierarchical and repeated measures nature of the data will be applied where necessary. The effect of soil and physiographic factors will be investigated. Alternative model formulations from the literature and those that might be developed during this research will be tested and those with the lowest prediction bias and variance and exhibit logical behavior will be adopted. Performance of the model system will be evaluated using 'leave-one-cluster-out' cross validation techniques and using sensitivity and uncertainty analysis techniques.Objective 2: Individual tree models for plantation loblolly and slash pine and even-aged natural longleaf pineDataThe same data used for Objective 1 will be used in this objective.AnalysisThe analysis will follow the same approach as described for Objective 1 except that the focus here will be on models that operate on individual trees.Objective 3: Establish in managed natural stands of red-oak sweetgum bottomland hardwoods and in LMAV hardwood plantationsThis objective will provide a set of long term study plots that will be measured at the time of plot establishment and re-measured at intervals of three to five years in the future. Data from these plots will eventually be used to develop growth and yield models for managed and unmanaged hardwood plantations in the LMAV and to develop thinning and stand improvement treatment response functions for the recently developed unmanaged natural red oak-sweetgum bottomland hardwoods model (Howard 2011; Jeffreys 2014).A total of at least 450 plots will be established, which will be distributed as: at least 150 in the managed natural red-oak sweetgum minor stream bottom stands, at least 150 in the unmanaged LMAV plantations, and a minimum of 150 in managed LMAV plantations. In the natural red-oak sweetgum minor stream bottom stands, the plots will be required to have hardwoods making up at least 70% of the total basal area with that of red oaks being at least 30% of the total. For the LMAV plantations, the species to be considered will depend on interest of the stakeholders and may include two-species mix plantations e.g. cottonwood-Nuttall oak (Quercus nuttallii Palmer) mixed species stands or single species stands. In the managed stands, each research plot location will have 4 plots - an unthinned control and 3 thinning, or related treatment, plots of light, moderate, and heavy thinning or related treatment intensities. The plots for each stand type will be selected to represent the widest possible range of age, stand density, and site quality classes. Circular plots will be used and plot size will vary between one-tenth of an acre and a size large enough (to a maximum of 1 acre) to capture at least 50 trees. Regeneration/Ingrowth data will be collected from one-hundredth acre plots within each research plot.On each main research plot, all trees with a diameter breast height (dbh) of at least 3.5 inches will be numbered, tagged, and the following data recorded for each tree:1. species,2. dbh,3. crown class,4. damage codes (if applicable, e.g., cankered, split, lightning damaged, stem rot), and5. distance and magnetic azimuth from plot center.On a representative sub sample of trees, additional measurements will be taken including:1. total tree height,2. height to base of the live crown, and3. crown width.On all trees containing 1, 2, or 3 factory sawlog(s), the following additional information will be collected:1. Length of each sawlog, and2. USDA Forest Service specifications sawlog grade.On the one-hundredth acre regeneration plot, seedlings will be numbered, tagged, and the following data recorded for each one of them:1. species,2. height, and3. seedling condition (if applicable, e.g., diseased, damaged).Objective 4: Propose new ways of accounting for genetic improvement effects in loblolly pine growth and yield models and incorporate some of the approaches in the models developed in Objectives 1 and 2DataStand growth and development data from genetic variety block plots experiments will be used to achieve this objective. Data will be obtained from ongoing studies, and from yet to be established studies, through formal and informal agreements or collaborations with other scientists at MSU and other universities in the region and with forest industry. Currently, informal agreements with other scientists at MSU are in place to use re-measurement data from a 29-year old half-sib genetic family by planting density study at MSU's John Starr Memorial Forest and also from a 7-year old clonal genetic variety by initial planting density by management intensities in Newton County, Mississippi. Other suitable data will be sought and incorporated in the research as they become available.AnalysisThe first stage of the analysis will involve an evaluation of the prediction behavior of the models in the loblolly pine growth and yield simulators currently available for use in the southern United States. Simulators to be considered include FASTLOB, PTAEDA, CLob., and the North Carolina State University Managed Pine Plantation Growth and Yield Simulator (Hafley and Smith 1989). The outputs of the whole model and the various sub-models of a simulator will be compared, across time, for different levels of genetic improvement effects that get incorporated in the simulator as site index change. In the second stage of the analysis, the degree of agreement between the changes predicted by the simulators and those observed from experimental plots data will be evaluated. Non-parametric statistical techniques will be used in this evaluation. Information on the nature and frequency of any inconsistencies between model predicted and observed changes will be used to formulate robust approaches of accounting for genetic improvement effects in loblolly pine growth and yield models and to suggest aspects of genetic improvement effects on growth and yield principles that might need to be investigated through experimentation. The formulated approaches will be incorporated into the loblolly pine models developed in Objectives 1 and 2.Objective 5: Propose the best statistical modeling approaches and model forms for the forest types in Mississippi and the mid Gulf region.This objective will be achieved alongside Objectives 1, 2, and 4. The modeling activities in these projects will involve evaluation of both existing and new modeling approaches to identify those that perform best. The best performing modeling approaches and model forms will be identified and recommended for use in the mid Gulf region.Objective 6: Develop new Windows® computer graphic and web-based user interface for the models developed in objectives 1 and 2The Microsoft Visual Studio® C++ programing platform will be used to integrate the models from Objectives 1 and 2 into Windows® application growth and yield simulators that will be availed as installable programs and also as web-based user interfaces.

Progress 01/12/15 to 02/16/18

Outputs
Target Audience: Nothing Reported Changes/Problems:Current PI left employment at Mississippi State University on August 11, 2017. The project will have to change. What opportunities for training and professional development has the project provided?One graduate completed a Masters degree, A new graduate student started work on the project in August 2017. Two undergraduate students received research experience training under the project. How have the results been disseminated to communities of interest?One oral presentation was given at the 2016 Joint Northeast and Southern Mensurationists Meeting that was held in Cooncord Massachusetts in November 2016. Four poster presentations were made at the 19th Biennial Southern Silvilcultural Conference that was held in Blacksburg Virginia in March 2017. Three poster presentations were made at University undergraduate and graduate student research symposia on Missississippi State University campus. What do you plan to do during the next reporting period to accomplish the goals? Nothing Reported

Impacts
What was accomplished under these goals? Objective 1: Develop new stand level diameter distribution models for plantation loblolly and slash pine and even-aged natural longleaf pine Models for plantation loblolly pine were developed. For natural longleaf pine, modeling data were gathered but no models were developed. Nothing was accomplished for slash pine. The following are the details of the accomplishments under the work on plantation loblolly pine models: A graduate student, name Binayak Bartaula, was recruited in Fall of 2015. He completed the development of stand level height growth, diameter growth, and mortality models for loblolly pine in mid-Gulf region of southern United States and completed a Masters Thesis based on the project. His equations were published as a proceedings publication titled "Equations for predicting stand level growth and survival for cut-over loblolly pine plantations in the mid-Gulf region of southern United States" in Kirschman, Julia E., comp. 2018. Proceedings of the 19th Biennial Southern Silvicultural Research Conference. e-Gen. Tech. Rep. SRS-234. Asheville, NC: U.S. Department of Agriculture Forest Service, Southern Research Station. 444 p. Binayak Bartaula graduated in July 2017. A student, name Josh Bankston, was recruited first as a Research Experience for Undergraduates student in Fall of 2016 then later as a Graduate Student in Fall of 2017 to investigate the various approaches of recovering stand diameter distribution from predictions of loblolly pine stand level diameter distribution models. His work was published as a proceedings publication titled "Plot size and prediction model form effects on stand diameter distribution recovery methods" in Kirschman, Julia E., comp. 2018. Proceedings of the 19th Biennial Southern Silvicultural Research Conference. e-Gen. Tech. Rep. SRS-234. Asheville, NC: U.S. Department of Agriculture Forest Service, Southern Research Station. 444 p. Josh Bankston was on schedule to graduate with a MS degree in Spring 2019. Objective 2: Develop individual tree models for plantation loblolly and slash pine and even-aged natural longleaf pine Data for developing loblolly and longleaf pine models were gathered cleaned and organized. Development of the models for the two species was not accomplished. No accomplishment on slash pine models. Objective 3: Establish new growth and yield plots in managed natural stands of red oak-sweetgum bottomland hardwoods and in Lower Mississippi Alluvial Valley hardwood plantations New growth and yield plots in natural red oak-sweetgum stands had not been established so no accomplishment in this area. For the Lower Mississippi Alluvial Valley, a 6-year Cooperative Agreement was signed with the US Forest Service in September 2016 to measure to analyze tree and biomass data from earlier-established research plantations of clonal cottonwood and willow. Objective 4: Propose new ways of accounting for genetic improvement effects in loblolly pine growth and yield models and incorporate some of the approaches in the models developed in Objectives 1 and 2 A study to investigate the effect of genetic similarity, of loblolly pine clones, in inter-tree competition models was carried out under a fixed price research agreement with the Forest Modeling Research Cooperative at Virginia Tech. Age 16 data were collected from established clonal and non-clonal loblolly pine study stands in coastal South Carolina in Spring of 2016 and analyzed. Based on the findings of this study, it was concluded that inter-tree competitive interactions did not affect diameter growth in clonal stands differently compared to non-clonal stands. Thus, it was not necessary to find new ways of accounting for genetic improvement effects in inter-tree competition model components of loblolly pine growth and yield models. Results of the analysis were published as a proceedings paper titled: Sixteen-year stand level growth and development of varietal and non-varietal loblolly pine in the Atlantic coastal plain of South Carolina in Kirschman, Julia E., comp. 2018. Proceedings of the 19th Biennial Southern Silvicultural Research Conference. e-Gen. Tech. Rep. SRS-234. Asheville, NC: U.S. Department of Agriculture Forest Service, Southern Research Station. 444 p. Objective 5: Propose the best statistical modeling approaches and model forms for the forest types in Mississippi and the mid-Gulf region. A few best modeling approaches were proposed. Based on the results of the work by the graduate students under Objective 1, it was proposed that: Physiographic region should be included in loblolly pine stand level tree mortality/survival models The percentile approach is the best approach of recovering loblolly pine stand's tree diameter distribution from stand summary characteristics predicted by stand level growth and yield models. Data for developing diameter distribution recovery system would need to be collected from study plots of size 0.2 acres. Larger plots would be unnecessarily too big and smaller plots would result in less accurate predictions. Objective 6: Develop new Windows® computer graphic and web-based user interface for the models developed in Objectives 1 and 2 No accomplishment

Publications


    Progress 10/01/15 to 09/30/16

    Outputs
    Target Audience: Nothing Reported Changes/Problems: Nothing Reported What opportunities for training and professional development has the project provided?One Masters level graduate student, Binayak Bartaula, completed the preparation of his Masters Thesis research plan and is on schedule in as far as his Master of Science program is concerned. Jacob Beard, a PhD who graduated in May 2016, incorporated a component of Objective 2 into his dissertation research. He developed an individual tree growth growth and mortality equations for loblolly pine. An undergraduate student, Joshua Bankston, was incorporated in to the project as a Research Experience for Undergraduates (REU) student. He began research on diameter distribution recovery of the loblolly component of Objective 1. How have the results been disseminated to communities of interest? Nothing Reported What do you plan to do during the next reporting period to accomplish the goals? Nothing Reported

    Impacts
    What was accomplished under these goals? Graduate student Binayak Bartaula worked on the loblolly pine stand level distribution model under Objective 1. His focus is on development of dominant height, diameter, and mortality prediction models. He completed the writing of his Masters Thesis research plan and will be carrying out data analysis and model development during the 2017 fiscal year. An undergraduate research scholar, Josh Bankston, started researching on the best way to recover predicted-stand diameter distribution from predicted stand level charactersistics. I researched on the role of upper-stem diameters in the taper equation volume prediction system of the loblolly pine model in Objective 1. A PhD student, Jacob Beard, who was supervised by the co-project directors and graduated in May 2016, developed a loblolly pine individual tree model under Objective 2 as part of his dissertation research. Under Objective 4, a $20,000 data analysis research grant from Forest Modeling Research Cooperative at Virginia Tech was secured for use in modeling competition in clonal loblolly pine stands. Data analysis under this grant was started and will continue in fiscal year 2017. A pre-proposal for a cooperative agreement with the USDA Forest Service, that will see the establishment of the measurment plots under Objective 3, was started.

    Publications


      Progress 01/12/15 to 09/30/15

      Outputs
      Target Audience: Nothing Reported Changes/Problems: Nothing Reported What opportunities for training and professional development has the project provided?One Masters level graduate student, name Binayak Bartaula,was recruited to work on the development of the loblolly pinediameter distribution stand levelmodel.An internal agreement to enjoin a current PhD student, name Jacob Beard,on the individual tree model development portion of the project was reached. How have the results been disseminated to communities of interest? Nothing Reported What do you plan to do during the next reporting period to accomplish the goals? Nothing Reported

      Impacts
      What was accomplished under these goals? Under Objectives 1 and 2, tree dimension data for the development of the loblolly pine diameter distribution model and those for the development of theloblolly pine individual tree model have been gathered, cleaned and organized into a format that can be analyzed.Extensive literature review on contemporary methods of modeling tree and stand development has also been done.Longleaf and slash pine data has been acquired. Under Objective 4, part of the data for modeling genetic effects on loblolly pine growth and yield have been acquired. Under Objective 3, an internal collaboration has been established with a fellow faculty member who hasan ongoing research in the Lower Mississippi Alluvial Valley hardwood plantations. Under this collaboration, ways of incorporating the growth and yield study in the ongoing study will be explored.

      Publications