Expert meeting on tree volume biomass allometric equations
Expert meeting on tree volume, biomass allometric equations Manual for tree allometric equations Authors : Saint-André L. (INRA-CIRAD), Picard N. (CIRAD), Henry M. (FAO), Sola G. (FAO) May 26, 2014 KFRI, INDIA
What is Allometry ? û Broad definition : within a given population, there is a statistical relationship between the size of an organism and the size of any part of it (Gould, 1966) CD For example: between height and diameter; diameter and crown size; biomass and diameter; etc. . û H D Can be used to predict some difficultto-measure tree characteristics from easily collected data. Volume prediction Ä Biomass prediction Nutrient content prediction Volume tables Ä Biomass equations Ä Nutrient content equations
What is Allometry ? û More restrictive definition : proportionality between the relative increments of two metrics measured on an organism (Huxley, 1924) relative increment in Biomass Which gives by integration Allometric coefficient relative increment in Diameter And by extension Where a gives the proportionality between the relative increments, b gives the proportionality between biomass and diameter (given a) and c is the biomass of the tree when D=0 (if D was measured at a height different from zero)
From the idea…. (in 2005) û Students are not familiar with the appropriate and up-to-date fitting techniques û the “magic” R 2 is usually preferred to the biological meaning of the equations û Models are fitted without considering the structure in the data set (source of variations) û Outliers are too easily removed from the data set while they can bring information on the structure of the dataset And so on …. There was then a strong need to make a new review on the methods to build tree allometric equations Including biological concepts, up-to-date statistical procedures and training examples
……To the result (2012) Picard N. , Saint-André L. , Henry M. 2012. Manuel de construction d’équations allométriques pour l’estimation du volume et la biomasse des arbres: de la mesure de terrain à la prédiction. Organisation des Nations Unies pour l’alimentation et ’agriculture, et Centre de Coopération Internationale en Recherche Agronomique pour le Développement, Rome, Montpellier, 222 pp. © 2012, CIRAD et FAO Available in French, Engish and Spanish http: //www. fao. org/forestry/fma/80797/en/
Content, 7 steps G. Validation of the model and prediction of trees volume or biomass A. Selection of trees biomass explanatory variables and the area of validity of the equation B. Design of the sampling and selection of the trees to be measured C. Preparation and implementatio n of field and laboratory measurements F. Fitting of the allometric equation E. Graphic exploration of the dataset D. Data entering and shaping
1 Tree growth, biomass partitioning and biomass allocation - Biological concepts Tree and stand growth: case of even-aged and monospecific forests -A- Stand production -Eichhorn’s rule -Assmann’s yield level theory -Langsaeter Hypothesis -Wood production (volume) of a given tree species at a given stand mean (or top) height should be identical for all site classes. - Soil fertility (site index) determines the time need to attain this height and volume. -There are some range of variations of wood production at a given top height (variations related to the stockability issue) -Losses in productivity if the standing stock is too low
2 Sampling strategy and stratification
3 Field work Step 2 û Above-ground biomasses Mixing leaves before taking a sample Sampling of cross-sections regularly along the trunk. The width of the crosssections should be fixed for all heights within the trees For multi-stem trees, take cross-sections in all stems Stump: Top limit = above-ground point where the tree was cut Down limit = where the roots could be clearly individualized Example, Rubber tree in Thaïland
4 Data recording and checking
5 Data analysis and graphical exploration of the structure in the data sets
6 Model fitting
5 Use of harmonised definitions of tree components Bg T Circumference or diameter (cm) at 1. 3 m Basal circumference or diameter (cm) S Rm Rf Rb Log volume (m 3) B Tree volume (m 3) F L Bt Leaf volume (m 3) Crown area (m 2) Branch volume (m 3) Crown height (m) Bd Log height (m) Tree height (m) Crown diameter (m) Basal area (m 2) B Bd Bg Bt L Rb Rf Rm S T F Bark Dead branches Gross branches: D>7 cm Thin branches: D<7 cm Leaves Large roots Fine roots Medium roots Stump Trunk-underbark Fruit/seed
7 Model use and biomass prediction
Training courses û û Fourth set in Sri Lanka after Ecuador (Ecuador, Panama, Paraguay), Vietnam (Vietnam and Indonesia) and Zambia (Zambia and Tanzania) 107 persons from more than 25 institutions
Objectives û to present the current and up-to-date knowledge for building allometric equations including courses on the related theory, field operations, fitting and use of the allometric equations, û to propose technical exercises aiming at identifying gaps (knowledge, allometric equations and raw data) to report carbon stocks and carbon stock changes at the country level, û to propose practical works on raw data to get familiar with the statistical software and to build allometric equations from their own data sets, û to help participants in elaborating their road map at national level (general scheme illustrating the required information and data fluxes from the forest stratification to the carbon stock assessments, identification of the existing raw data and allometric equations, bringing to light the gaps, quality control procedures, elaboration of a preliminary road map), û to elaborate a translational network of experts on allometric equations.
Content Identification of the data to be used for the exercise Use and prediction Practical cases and issues related to model fitting G. Validation of the model and prediction of trees volume or biomass A. Selection of trees biomass explanatory variables and the area of validity of the equation Complexity of tree growth and biomass allocation B. Design of the sampling and selection of the trees to be measured C. Preparation and implementatio n of field and laboratory measurements F. Fitting of the allometric equation E. Graphic exploration of the dataset Graphical analysis and first data interpretation D. Data entering and shaping Sampling strategy and stratification Field work: example from case studies Data entering, data management and QC
Planning Resource persons/facilitator Presentations DAY 1: Overview of the current status of the development of allometric equations in the countries Presentation of the content of the training Forest types and ecological stratification of Vietnam (Including presentation on the structure of the forest based on data from previous NFI cycles) and AE development plan under UN-REDD Vietnam Current status of knowledge on volume and biomass assessment (including description of the database, methodology used, sources, quality of the raw data, wood density, etc. ) Experiences from destructive measurement field work and identified gaps overall Matieu https: //dl. dropbox. com/u/6896677/Content%20 of%20 the%2 0 training-web. pdf Vietnam Study Team https: //dl. dropbox. com/u/6896677/2012_VTP_Forests%20 a (including 20 min presentation nd%20 AE%20 development. pdf by Phuong, RCFEE) Forest types and ecological stratification of Indonesia (Including presentation on the structure of the forest based on data from previous NFI cycles) Indonesia Vietnam Study Team https: //dl. dropbox. com/u/6896677/Current%20 status%20 of (including 20 min presentation %20 knowledge%20 on%20 volume%20 and%20 biomass. pdf by Hung, FIPI) Vietnam (including 10 min *3 presentation from VFU, TNU, NW-Sub-FIPI) VFU: https: //dl. dropbox. com/u/6896677/VFU_Experiences 2. pdf TNU: https: //dl. dropbox. com/u/6896677/Experiences%20 from%2 0 destructive%20 measurement%20 allometric%20 equations% 20 development%20 -%20 TNU%202. pdf Sub-FIPI: https: //dl. dropbox. com/u/6896677/Experience%20 and%20 p roposal%20 for%20 AE%20 -%20 NW%20 Sub%20 FIPI. pdf https: //dl. dropbox. com/u/6896677/presentation%20 in%20 Vi etnam-revised. pdf
Planning DAY 2: Development of allometric equations “state of the art” Building tree allometric equation to assess volume and biomass : from the field to the prediction Complexity of tree growth and biomass allocation Sampling strategy and stratification 1. Which are the forest stratum without AE in your country? 2. Do those forest stratums contribute significantly to emissions? 3. What is the max and min tree diameter sampled? How many tree sample per diameter class? Field work: example from case studies Data entering, data management and QC Graphical analysis and first data interpretation Practical cases and issues related to model fitting Use and prediction Matieu Laurent discussions Laurent / Matieu Laurent https: //dl. dropbox. com/u/6896677/Building%20 tree%20 all ometric%20 equation%20 to%20 assess%20 volume%20 and% 20 biomass%20%5 BAutosaved%5 D. pdf https: //dl. dropbox. com/u/6896677/Biomass_Training. Mate rial_From. Biology. To. AE. pdf https: //dl. dropbox. com/u/6896677/Biomass_Training. Mate rial_Sampling. Strategy. pdf VN table 1: https: //dl. dropbox. com/u/6896677/Presentation_Table_1 _Day 2. pdf VN table 2: https: //dl. dropbox. com/u/6896677/Presentation. pdf VN TNU: https: //dl. dropbox. com/u/6896677/AE%20 and%20 N%20%20 TNU. pdf Indon 1: https: //dl. dropbox. com/u/6896677/Indon. Question%202. pdf https: //dl. dropbox. com/u/6896677/Field%20 work_exampl e%20 from%20 case%20 studies%20%5 BAutosaved%5 D. pdf https: //dl. dropbox. com/u/6896677/Data%20 entering%2 C %20 data%20 management%20 and%20 QC. pdf https: //dl. dropbox. com/u/6896677/Graphical%20 analysis %20 and%20 first%20 data%20 interpretation. pdf https: //dl. dropbox. com/u/6896677/Biomass_Training. Mate rial_Model. Fitting. pdf https: //dl. dropbox. com/u/6896677/Biomass_Training. Mate rial_Model. Use. pdf
Planning DAY 3 -4: Exercises on development of allometric models Installation of the statistical By group software Graphic exploration (different set of data) Model fitting with and without heteroscedasticity by tree compartment Model fitting for total aboveground biomass (with or without additivity) Biomass assessment using existing data and models (development of decisional trees) 1. Which are the different equation forms (the main commons)? 2. Which are the inputs in your equations? 3. Which are the statistical information you collected to describe those equations? 1. Have you already used non linear model fitting methods 2. Can you provide a concrete case where the R 2 does not indicate the best model? 3: Can you explain the different steps for model development and selection? SAS software installed per group (for the training only) Exercise using sample data (exercise done on database) Indonesia 1: https: //dl. dropbox. com/u/6896677/Answers%20 Group%20 1 -Model%20 fitting. docx VN 1: https: //dl. dropbox. com/u/6896677/PP_Huong. pptx
- Slides: 20